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

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

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

code/trunk/pcre_internal.h revision 483 by ph10, Mon Jan 4 16:19:17 2010 UTC code/branches/pcre16/pcre_internal.h revision 764 by zherczeg, Wed Nov 23 17:23:20 2011 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2010 University of Cambridge             Copyright (c) 1997-2011 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 40  POSSIBILITY OF SUCH DAMAGE. Line 40  POSSIBILITY OF SUCH DAMAGE.
40    
41  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_" or "_pcre16_" depending on
44    the PRIV macro. */
45    
46  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
47  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
# Line 51  functions whose names all begin with "_p Line 52  functions whose names all begin with "_p
52  #define PCRE_DEBUG  #define PCRE_DEBUG
53  #endif  #endif
54    
55  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"  /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56    #ifndef COMPILE_PCRE16
57    #define COMPILE_PCRE8
58    #endif
59    
60    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
61  script prevents both being selected, but not everybody uses "configure". */  script prevents both being selected, but not everybody uses "configure". */
62    
63  #if defined EBCDIC && defined SUPPORT_UTF8  #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
64  #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.  #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
65  #endif  #endif
66    
67  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
# Line 158  set, we ensure here that it has no effec Line 164  set, we ensure here that it has no effec
164  #define PCRE_CALL_CONVENTION  #define PCRE_CALL_CONVENTION
165  #endif  #endif
166    
167  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
168  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
169  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
170  systems. Instead we make use of the maximum sizes that are available at  systems. Instead we make use of the maximum sizes that are available at
171  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
172    
173    typedef unsigned char pcre_uint8;
174    
175  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
176    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
177    typedef short pcre_int16;    typedef short pcre_int16;
# Line 184  preprocessor time in standard C environm Line 192  preprocessor time in standard C environm
192    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
193  #endif  #endif
194    
195  /* When checking for integer overflow in pcre_compile(), we need to handle  /* When checking for integer overflow in pcre_compile(), we need to handle
196  large integers. If a 64-bit integer type is available, we can use that.  large integers. If a 64-bit integer type is available, we can use that.
197  Otherwise we have to cast to double, which of course requires floating point  Otherwise we have to cast to double, which of course requires floating point
198  arithmetic. Handle this by defining a macro for the appropriate type. If  arithmetic. Handle this by defining a macro for the appropriate type. If
199  stdint.h is available, include it; it may define INT64_MAX. The macro int64_t  stdint.h is available, include it; it may define INT64_MAX. Systems that do not
200  may be set by "configure". */  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
201    by "configure". */
202    
203  #if HAVE_STDINT_H  #if HAVE_STDINT_H
204  #include <stdint.h>  #include <stdint.h>
205    #elif HAVE_INTTYPES_H
206    #include <inttypes.h>
207  #endif  #endif
208    
209  #if defined INT64_MAX || defined int64_t  #if defined INT64_MAX || defined int64_t
# Line 203  may be set by "configure". */ Line 214  may be set by "configure". */
214    
215  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
216  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
217  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char * or short *,
218  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. We define a short type
219  to save lots of typing. I tried "uchar", but it causes problems on Digital  for the current character representation (either 8 or 16 bit) to save lots
220  Unix, where it is defined in sys/types, so use "uschar" instead. */  of typing. I tried "uchar", but it causes problems on Digital Unix, where
221    it is defined in sys/types, so use "uschar" instead. */
222    
223  typedef unsigned char uschar;  #ifndef COMPILE_PCRE16
224    typedef unsigned char pcre_uchar;
225    #else
226    #if USHRT_MAX != 65535
227    /* This is a warning message. Change PCRE_SCHAR16 to a 16 bit data type in
228    pcre.h(.in) and disable (comment out) this message. */
229    #error Warning: PCRE_SCHAR16 is not a 16 bit data type.
230    #endif
231    typedef pcre_uint16 pcre_uchar;
232    #endif
233    
234  /* This is an unsigned int value that no character can ever have. UTF-8  /* This is an unsigned int value that no character can ever have. UTF-8
235  characters only go up to 0x7fffffff (though Unicode doesn't go beyond  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
# Line 231  start/end of string field names are. */ Line 252  start/end of string field names are. */
252  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
253    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
254      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
255       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
256         utf8)) \         utf8)) \
257      : \      : \
258      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
# Line 245  start/end of string field names are. */ Line 266  start/end of string field names are. */
266  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
267    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
268      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
269       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
270         &(NLBLOCK->nllen), utf8)) \         &(NLBLOCK->nllen), utf8)) \
271      : \      : \
272      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
# Line 264  used for the external interface and appe Line 285  used for the external interface and appe
285  must begin with PCRE_. */  must begin with PCRE_. */
286    
287  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
288  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
289  #else  #else
290  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
 #endif  
   
291    
292    /* PCRE_SPTR is defined in pcre.h. */
293    #define USPTR const uschar *
294    #endif
295    
296  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
297  values. */  values. */
# Line 340  The macros are controlled by the value o Line 360  The macros are controlled by the value o
360  the config.h file, but can be overridden by using -D on the command line. This  the config.h file, but can be overridden by using -D on the command line. This
361  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
362    
363    #ifdef COMPILE_PCRE8
364    
365  #if LINK_SIZE == 2  #if LINK_SIZE == 2
366    
367  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 376  is automated on Unix systems via the "co Line 398  is automated on Unix systems via the "co
398  #define GET(a,n) \  #define GET(a,n) \
399    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
400    
401  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
402    #define MAX_PATTERN_SIZE (1 << 30)
403    
404    #else
405    #error LINK_SIZE must be either 2, 3, or 4
406    #endif
407    
408    #else /* COMPILE_PCRE8 */
409    
410    #ifdef COMPILE_PCRE16
411    
412    #if LINK_SIZE == 2
413    
414    #undef LINK_SIZE
415    #define LINK_SIZE 1
416    
417    #define PUT(a,n,d)   \
418      (a[n] = (d))
419    
420    #define GET(a,n) \
421      (a[n])
422    
423    #define MAX_PATTERN_SIZE (1 << 16)
424    
425    #elif LINK_SIZE == 3 || LINK_SIZE == 4
426    
427    #define PUT(a,n,d)   \
428      (a[n] = (d) >> 16), \
429      (a[(n)+1] = (d) & 65536)
430    
431    #define GET(a,n) \
432      (((a)[n] << 16) | (a)[(n)+1])
433    
434    /* Keep it positive */
435    #define MAX_PATTERN_SIZE (1 << 30)
436    
437  #else  #else
438  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
439  #endif  #endif
440    
441    #else
442    #error Unsupported compiling mode
443    #endif /* COMPILE_PCRE16 */
444    
445    #endif /* COMPILE_PCRE8 */
446    
447  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
448    
# Line 393  is automated on Unix systems via the "co Line 453  is automated on Unix systems via the "co
453  offsets changes. There are used for repeat counts and for other things such as  offsets changes. There are used for repeat counts and for other things such as
454  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
455    
456    #ifdef COMPILE_PCRE8
457    
458    #define IMM2_SIZE 2
459    
460  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
461    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
462    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 400  capturing parenthesis numbers in back re Line 464  capturing parenthesis numbers in back re
464  #define GET2(a,n) \  #define GET2(a,n) \
465    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
466    
467  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #else /* COMPILE_PCRE8 */
468    
469    #ifdef COMPILE_PCRE16
470    
471    #define IMM2_SIZE 1
472    
473    #define PUT2(a,n,d)   \
474       a[n] = d
475    
476    #define GET2(a,n) \
477       a[n]
478    
479    #else
480    #error Unsupported compiling mode
481    #endif /* COMPILE_PCRE16 */
482    
483    #endif /* COMPILE_PCRE8 */
484    
485    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
486    
487  /* 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
488  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
489  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
490  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
491  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
492    UTF-8 support is omitted, we don't even define them. */
493    
494  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
495  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 415  support is omitted, we don't even define Line 497  support is omitted, we don't even define
497  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
498  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
499  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
500    /* #define GETCHARLENTEST(c, eptr, len) */
501  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
502    
503  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
504    
505    /* These macros were originally written in the form of loops that used data
506    from the tables whose names start with PRIV(utf8_table). They were rewritten by
507    a user so as not to use loops, because in some environments this gives a
508    significant performance advantage, and it seems never to do any harm. */
509    
510    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
511    advancing the pointer. */
512    
513    #define GETUTF8(c, eptr) \
514        { \
515        if ((c & 0x20) == 0) \
516          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
517        else if ((c & 0x10) == 0) \
518          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
519        else if ((c & 0x08) == 0) \
520          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
521          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
522        else if ((c & 0x04) == 0) \
523          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
524              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
525              (eptr[4] & 0x3f); \
526        else \
527          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
528              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
529              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
530        }
531    
532  /* 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
533  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
534    
535  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
536    c = *eptr; \    c = *eptr; \
537    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; \  
       } \  
     }  
538    
539  /* 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
540  pointer. */  pointer. */
541    
542  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
543    c = *eptr; \    c = *eptr; \
544    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
545    
546    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
547    the pointer. */
548    
549    #define GETUTF8INC(c, eptr) \
550      { \      { \
551      int gcii; \      if ((c & 0x20) == 0) \
552      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
553      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
554      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
555      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
556          eptr += 2; \
557          } \
558        else if ((c & 0x08) == 0) \
559        { \        { \
560        gcss -= 6; \        c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
561        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
562          eptr += 3; \
563          } \
564        else if ((c & 0x04) == 0) \
565          { \
566          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
567              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
568              (eptr[3] & 0x3f); \
569          eptr += 4; \
570          } \
571        else \
572          { \
573          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
574              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
575              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
576          eptr += 5; \
577        } \        } \
578      }      }
579    
# Line 460  know we are in UTF-8 mode. */ Line 582  know we are in UTF-8 mode. */
582    
583  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
584    c = *eptr++; \    c = *eptr++; \
585    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; \  
       } \  
     }  
586    
587  /* 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.
588    This is called when we don't know if we are in UTF-8 mode. */
589    
590  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
591    c = *eptr++; \    c = *eptr++; \
592    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
593    
594    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
595    advancing the pointer, incrementing the length. */
596    
597    #define GETUTF8LEN(c, eptr, len) \
598      { \      { \
599      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
600      int gcss = 6*gcaa; \        { \
601      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
602      while (gcaa-- > 0) \        len++; \
603          } \
604        else if ((c & 0x10)  == 0) \
605          { \
606          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
607          len += 2; \
608          } \
609        else if ((c & 0x08)  == 0) \
610          {\
611          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
612              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
613          len += 3; \
614          } \
615        else if ((c & 0x04)  == 0) \
616        { \        { \
617        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
618        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
619              (eptr[4] & 0x3f); \
620          len += 4; \
621          } \
622        else \
623          {\
624          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
625              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
626              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
627          len += 5; \
628        } \        } \
629      }      }
630    
# Line 493  if there are extra bytes. This is called Line 633  if there are extra bytes. This is called
633    
634  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
635    c = *eptr; \    c = *eptr; \
636    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
637    
638  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
639  pointer, incrementing length if there are extra bytes. This is called when we  pointer, incrementing length if there are extra bytes. This is called when we
640  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
641    
642  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
643    c = *eptr; \    c = *eptr; \
644    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
645    
646  /* 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
647  it is. This is called only in UTF-8 mode - we don't put a test within the macro  it is. This is called only in UTF-8 mode - we don't put a test within the macro
# Line 533  because almost all calls are already wit Line 649  because almost all calls are already wit
649    
650  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
651    
652  #endif  #endif  /* SUPPORT_UTF8 */
653    
654    
655  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
# Line 544  Standard C system should have one. */ Line 660  Standard C system should have one. */
660  #endif  #endif
661    
662    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
663  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
664  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
665  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
# Line 560  compatibility. */ Line 672  compatibility. */
672  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
673  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
674  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
675    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
676    
677  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
678    
679  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
680  #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */  #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
681    
682  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
683  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 577  time, run time, or study time, respectiv Line 690  time, run time, or study time, respectiv
690     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
691     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| \
692     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
693     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
694    
695  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
696    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
# Line 590  time, run time, or study time, respectiv Line 703  time, run time, or study time, respectiv
703     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
704     PCRE_NO_START_OPTIMIZE)     PCRE_NO_START_OPTIMIZE)
705    
706  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
707       PCRE_STUDY_JIT_COMPILE
708    
709  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. Also used
710  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 872  so that PCRE works on both ASCII and EBC Line 986  so that PCRE works on both ASCII and EBC
986  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
987  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
988  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
989    #define STRING_MARK0                "MARK\0"
990  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
991  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
992  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 893  so that PCRE works on both ASCII and EBC Line 1008  so that PCRE works on both ASCII and EBC
1008    
1009  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
1010    
1011  #define STRING_CR_RIGHTPAR          "CR)"  #define STRING_CR_RIGHTPAR             "CR)"
1012  #define STRING_LF_RIGHTPAR          "LF)"  #define STRING_LF_RIGHTPAR             "LF)"
1013  #define STRING_CRLF_RIGHTPAR        "CRLF)"  #define STRING_CRLF_RIGHTPAR           "CRLF)"
1014  #define STRING_ANY_RIGHTPAR         "ANY)"  #define STRING_ANY_RIGHTPAR            "ANY)"
1015  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1016  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1017  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1018  #define STRING_UTF8_RIGHTPAR        "UTF8)"  #define STRING_UTF8_RIGHTPAR           "UTF8)"
1019    #define STRING_UCP_RIGHTPAR            "UCP)"
1020    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1021    
1022  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF8 */
1023    
# Line 1124  only. */ Line 1241  only. */
1241  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1242  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1243  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1244    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1245  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1246  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1247  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1145  only. */ Line 1263  only. */
1263    
1264  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1265    
1266  #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1267  #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1268  #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1269  #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1270  #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1271  #define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1272  #define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1273  #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1274    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1275    #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
1276    
1277  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
1278    
# Line 1185  only. */ Line 1305  only. */
1305    
1306  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1307  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1308  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1309  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1310  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1311    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1312    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1313    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1314    #define PT_WORD       8    /* Word - L plus N plus underscore */
1315    
1316  /* 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
1317  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 1204  contain UTF-8 characters with values gre Line 1328  contain UTF-8 characters with values gre
1328  /* 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
1329  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
1330  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
1331  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
1332  corresponds to "." rather than an escape sequence, and another for OP_ALLANY  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1333  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1334    non-DOTALL mode, "." behaves like \N.
1335    
1336    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1337    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1338    They must be contiguous, and remain in order so that the replacements can be
1339    looked up from a table.
1340    
1341  The final escape must be ESC_REF as subsequent values are used for  The final escape must be ESC_REF as subsequent values are used for
1342  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
# Line 1216  put in between that don't consume a char Line 1346  put in between that don't consume a char
1346  */  */
1347    
1348  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,
1349         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1350         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1351           ESC_E, ESC_Q, ESC_g, ESC_k,
1352           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1353         ESC_REF };         ESC_REF };
1354    
   
1355  /* 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
1356  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.
1357    
# Line 1244  enum { Line 1375  enum {
1375    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1376    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1377    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1378    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1379    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1380      OP_ALLANY,         /* 13 Match any character */
1381    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1382    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1383    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1258  enum { Line 1390  enum {
1390    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1391    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1392    
1393    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1394    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1395    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1396    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1397    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1398    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1399      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1400    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1401    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1402    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1403    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1404    OP_QUERY,          /* 35 */  
1405    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1406    
1407    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1408    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1409    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1410      OP_MINPLUS,        /* 36 */
1411    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1412    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1413    OP_POSQUERY,       /* 42 Posesssified query */  
1414    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1415      OP_MINUPTO,        /* 40 */
1416    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1417    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1418    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1419    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1420    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1421    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1422    
1423    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1424    OP_NOTMINUPTO,     /* 51 */  
1425    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1426      OP_MINSTARI,       /* 47 */
1427    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1428    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1429    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1430    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1431    
1432    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1433    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1434    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1435    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1436    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1437    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1438      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1439    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1440    OP_TYPEMINUPTO,    /* 64 */  
1441    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1442      /**** Negated single character, caseful; must precede the caseless ones ****/
1443    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1444    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1445    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1446    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1447      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1448    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1449    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1450    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1451    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1452    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1453    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1454    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1455    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1456      OP_NOTPOSPLUS,     /* 69 */
1457    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1458    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1459                             class - the difference is relevant only when a UTF-8  
1460                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1461    
1462    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1463                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1464      OP_NOTPLUSI,       /* 74 */
1465    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1466    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1467    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1468    
1469    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1470    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1471    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1472    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1473      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1474    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1475      OP_NOTPOSQUERYI,   /* 83 */
1476    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1477    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1478    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1479    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1480    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1481      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1482    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1483    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1484      OP_TYPEQUERY,      /* 89 */
1485    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1486    OP_BRA,            /* 94 Start of non-capturing bracket */  
1487    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1488    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1489      OP_TYPEEXACT,      /* 93 Exactly n matches */
1490    
1491      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1492      OP_TYPEPOSPLUS,    /* 95 */
1493      OP_TYPEPOSQUERY,   /* 96 */
1494      OP_TYPEPOSUPTO,    /* 97 */
1495    
1496      /* These are used for character classes and back references; only the
1497      first six are the same as the sets above. */
1498    
1499      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1500      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1501      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1502      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1503      OP_CRQUERY,        /* 102 */
1504      OP_CRMINQUERY,     /* 103 */
1505    
1506      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1507      OP_CRMINRANGE,     /* 105 */
1508    
1509      /* End of quantifier opcodes */
1510    
1511      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1512      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1513                                  class - the difference is relevant only when a
1514                                  UTF-8 character > 255 is encountered. */
1515      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1516                                  class. This does both positive and negative. */
1517      OP_REF,            /* 109 Match a back reference, casefully */
1518      OP_REFI,           /* 110 Match a back reference, caselessly */
1519      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1520      OP_CALLOUT,        /* 112 Call out to external function if provided */
1521    
1522      OP_ALT,            /* 113 Start of alternation */
1523      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1524      OP_KETRMAX,        /* 115 These two must remain together and in this */
1525      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1526      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1527    
1528      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1529      asserts must remain in order. */
1530    
1531      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1532      OP_ASSERT,         /* 119 Positive lookahead */
1533      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1534      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1535      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1536    
1537      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1538      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1539      subpattern that isn't an assertion. The POS versions must immediately follow
1540      the non-POS versions in each case. */
1541    
1542      OP_ONCE,           /* 123 Atomic group, contains captures */
1543      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1544      OP_BRA,            /* 125 Start of non-capturing bracket */
1545      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1546      OP_CBRA,           /* 127 Start of capturing bracket */
1547      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1548      OP_COND,           /* 129 Conditional group */
1549    
1550    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1551    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1552    
1553    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1554    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1555    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1556      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1557      OP_SCOND,          /* 134 Conditional group, check empty */
1558    
1559    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
1560    
1561    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 135 Used to hold a capture number as condition */
1562    OP_NCREF,          /* 101 Same, but generaged by a name reference*/    OP_NCREF,          /* 136 Same, but generated by a name reference*/
1563    OP_RREF,           /* 102 Used to hold a recursion number as condition */    OP_RREF,           /* 137 Used to hold a recursion number as condition */
1564    OP_NRREF,          /* 103 Same, but generaged by a name reference*/    OP_NRREF,          /* 138 Same, but generated by a name reference*/
1565    OP_DEF,            /* 104 The DEFINE condition */    OP_DEF,            /* 139 The DEFINE condition */
1566    
1567    OP_BRAZERO,        /* 105 These two must remain together and in this */    OP_BRAZERO,        /* 140 These two must remain together and in this */
1568    OP_BRAMINZERO,     /* 106 order. */    OP_BRAMINZERO,     /* 141 order. */
1569      OP_BRAPOSZERO,     /* 142 */
1570    
1571    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1572    
1573    OP_PRUNE,          /* 107 */    OP_MARK,           /* 143 always has an argument */
1574    OP_SKIP,           /* 108 */    OP_PRUNE,          /* 144 */
1575    OP_THEN,           /* 109 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1576    OP_COMMIT,         /* 110 */    OP_SKIP,           /* 146 */
1577      OP_SKIP_ARG,       /* 147 same, but with argument */
1578      OP_THEN,           /* 148 */
1579      OP_THEN_ARG,       /* 149 same, but with argument */
1580      OP_COMMIT,         /* 150 */
1581    
1582    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1583    
1584    OP_FAIL,           /* 111 */    OP_FAIL,           /* 151 */
1585    OP_ACCEPT,         /* 112 */    OP_ACCEPT,         /* 152 */
1586    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1587      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1588    
1589    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1590    
1591    OP_SKIPZERO        /* 114 */    OP_SKIPZERO,       /* 155 */
1592    
1593      /* This is not an opcode, but is used to check that tables indexed by opcode
1594      are the correct length, in order to catch updating errors - there have been
1595      some in the past. */
1596    
1597      OP_TABLE_LENGTH
1598  };  };
1599    
1600  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1601  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
1602  called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */  called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1603    
1604    
1605  /* 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
1606  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
1607    only in pcre_printint.c, which fills out the full names in many cases (and in
1608    some cases doesn't actually use these names at all). */
1609    
1610  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1611    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1612    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1613    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1614    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1615    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1616    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1617      "{", "{", "{",                                                  \
1618    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1619    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1620      "{", "{", "{",                                                  \
1621      "*+","++", "?+", "{",                                           \
1622      "*", "*?", "+", "+?", "?", "??",                                \
1623      "{", "{", "{",                                                  \
1624      "*+","++", "?+", "{",                                           \
1625      "*", "*?", "+", "+?", "?", "??",                                \
1626      "{", "{", "{",                                                  \
1627    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1628    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1629    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1630    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1631    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1632    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1633    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1634    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1635      "Once", "Once_NC",                                              \
1636      "Bra", "BraPos", "CBra", "CBraPos",                             \
1637      "Cond",                                                         \
1638      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1639      "SCond",                                                        \
1640    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1641    "Brazero", "Braminzero",                                        \    "Brazero", "Braminzero", "Braposzero",                          \
1642    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1643      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1644      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1645    "Close", "Skip zero"    "Close", "Skip zero"
1646    
1647    
# Line 1436  in UTF-8 mode. The code that uses this t Line 1659  in UTF-8 mode. The code that uses this t
1659    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1660    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1661    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1662    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1663    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1664    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1665      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1666    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1667    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1668    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1669    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1670    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1671    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1672      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1673    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1674      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1675      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1676      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1677    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1678    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1679    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1680    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1681      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1682      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1683      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1684    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1685    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1686    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 1461  in UTF-8 mode. The code that uses this t Line 1692  in UTF-8 mode. The code that uses this t
1692   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1693    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1694    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1695      3,                             /* REFI                                   */ \
1696    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1697    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1698    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1699    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1700    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1701    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1702      1+LINK_SIZE,                   /* KetRpos                                */ \
1703      1+LINK_SIZE,                   /* Reverse                                */ \
1704    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1705    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1706    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1707    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1708    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1709      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1710    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1711      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1712    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1713      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1714    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1715    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1716      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1717    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1718      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1719    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1720    3, 3,                          /* CREF, NCREF                            */ \    3, 3,                          /* CREF, NCREF                            */ \
1721    3, 3,                          /* RREF, NRREF                            */ \    3, 3,                          /* RREF, NRREF                            */ \
1722    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1723    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1724    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1725    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1726      1, 3,                          /* THEN, THEN_ARG                         */ \
1727      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1728      3, 1                           /* CLOSE, SKIPZERO  */
1729    
1730  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1731  condition. */  condition. */
1732    
1733  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1734    
1735  /* 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
1736  tracked. */  easily be tracked. When a new number is added, the table called eint in
1737    pcreposix.c must be updated. */
1738    
1739  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1740         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1501  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1742  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1742         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1743         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1744         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1745         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1746           ERR70, ERRCOUNT };
1747    
1748  /* 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
1749  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 1535  typedef struct real_pcre { Line 1777  typedef struct real_pcre {
1777    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1778    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1779    
1780    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1781    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1782  } real_pcre;  } real_pcre;
1783    
1784  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
# Line 1545  remark (see NOTE above) about extending Line 1787  remark (see NOTE above) about extending
1787  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1788    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1789    pcre_uint32 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
1790    uschar start_bits[32];          /* Starting char bits */    pcre_uint8 start_bits[32];      /* Starting char bits */
1791    pcre_uint32 minlength;          /* Minimum subject length */    pcre_uint32 minlength;          /* Minimum subject length */
1792  } pcre_study_data;  } pcre_study_data;
1793    
1794  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1795  compiling, so that instructions to close them can be compiled when (*ACCEPT) is  compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1796  encountered. */  encountered. This is also used to identify subpatterns that contain recursive
1797    back references to themselves, so that they can be made atomic. */
1798    
1799  typedef struct open_capitem {  typedef struct open_capitem {
1800    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1801    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1802      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1803  } open_capitem;  } open_capitem;
1804    
1805  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1806  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1807    
1808  typedef struct compile_data {  typedef struct compile_data {
1809    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1810    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1811    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1812    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1813    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1814    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1815    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1816    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1817    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;          /* Chain of open capture items */
1818    uschar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
1819    uschar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
1820    int  names_found;             /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
1821    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
1822    int  bracount;                /* Count of capturing parens as we compile */    int  bracount;                    /* Count of capturing parens as we compile */
1823    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
1824    int  top_backref;             /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
1825    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
1826    int  external_options;        /* External (initial) options */    int  assert_depth;                /* Depth of nested assertions */
1827    int  external_flags;          /* External flag bits to be set */    int  external_options;            /* External (initial) options */
1828    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_flags;              /* External flag bits to be set */
1829    BOOL had_accept;              /* (*ACCEPT) encountered */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1830    BOOL check_lookbehind;        /* Lookbehinds need later checking */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1831    int  nltype;                  /* Newline type */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1832    int  nllen;                   /* Newline string length */    int  nltype;                      /* Newline type */
1833    uschar nl[4];                 /* Newline string when fixed length */    int  nllen;                       /* Newline string length */
1834      pcre_uchar nl[4];                 /* Newline string when fixed length */
1835  } compile_data;  } compile_data;
1836    
1837  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1838  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1839    
1840  typedef struct branch_chain {  typedef struct branch_chain {
1841    struct branch_chain *outer;    struct branch_chain *outer;
1842    uschar *current_branch;    pcre_uchar *current_branch;
1843  } branch_chain;  } branch_chain;
1844    
1845  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1846  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1847    
1848  typedef struct recursion_info {  typedef struct recursion_info {
1849    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1850    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1851    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1852    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
1853    int *offset_save;             /* Pointer to start of saved offsets */    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
   int save_offset_top;          /* Current value of offset_top */  
1854  } recursion_info;  } recursion_info;
1855    
1856    /* A similar structure for pcre_dfa_exec(). */
1857    
1858    typedef struct dfa_recursion_info {
1859      struct dfa_recursion_info *prevrec;
1860      int group_num;
1861      PCRE_PUCHAR subject_position;
1862    } dfa_recursion_info;
1863    
1864  /* 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
1865  pointer at the start of each subpattern, so as to detect when an empty string  pointer at the start of each subpattern, so as to detect when an empty string
1866  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1867    pcre_exec(). */
1868    
1869  typedef struct eptrblock {  typedef struct eptrblock {
1870    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
1871    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
1872  } eptrblock;  } eptrblock;
1873    
1874    
# Line 1627  typedef struct match_data { Line 1879  typedef struct match_data {
1879    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
1880    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
1881    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
1882    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
1883    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
1884    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
1885    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
1886    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
1887    int    name_count;            /* Number of names in name table */    int    name_count;              /* Number of names in name table */
1888    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
1889    uschar *name_table;           /* Table of names */    pcre_uchar *name_table;         /* Table of names */
1890    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];               /* Newline string when fixed */
1891    const uschar *lcc;            /* Points to lower casing table */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
1892    const uschar *ctypes;         /* Points to table of type maps */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
1893    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;         /* Set if too many extractions */
1894    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                  /* NOTBOL flag */
1895    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                  /* NOTEOL flag */
1896    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                    /* UTF8 flag */
1897    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
1898    BOOL   endonly;               /* Dollar not before final \n */    BOOL   use_ucp;                 /* PCRE_UCP flag */
1899    BOOL   notempty;              /* Empty string match not wanted */    BOOL   endonly;                 /* Dollar not before final \n */
1900    BOOL   notempty_atstart;      /* Empty string match at start not wanted */    BOOL   notempty;                /* Empty string match not wanted */
1901    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
1902    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   hitend;                  /* Hit the end of the subject at some point */
1903    const uschar *start_code;     /* For use when recursing */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
1904    USPTR  start_subject;         /* Start of the subject string */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
1905    USPTR  end_subject;           /* End of the subject string */    const  pcre_uchar *start_code;  /* For use when recursing */
1906    USPTR  start_match_ptr;       /* Start of matched string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
1907    USPTR  end_match_ptr;         /* Subject position at end match */    PCRE_PUCHAR end_subject;        /* End of the subject string */
1908    USPTR  start_used_ptr;        /* Earliest consulted character */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
1909    int    partial;               /* PARTIAL options */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
1910    int    end_offset_top;        /* Highwater mark at end of match */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
1911    int    capture_last;          /* Most recent capture number */    int    partial;                 /* PARTIAL options */
1912    int    start_offset;          /* The start offset value */    int    end_offset_top;          /* Highwater mark at end of match */
1913    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    int    capture_last;            /* Most recent capture number */
1914    int    eptrn;                 /* Next free eptrblock */    int    start_offset;            /* The start offset value */
1915    recursion_info *recursive;    /* Linked list of recursion data */    int    match_function_type;     /* Set for certain special calls of MATCH() */
1916    void  *callout_data;          /* To pass back to callouts */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
1917      int    eptrn;                   /* Next free eptrblock */
1918      recursion_info *recursive;      /* Linked list of recursion data */
1919      void  *callout_data;            /* To pass back to callouts */
1920      const  pcre_uchar *mark;        /* Mark pointer to pass back */
1921      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
1922  } match_data;  } match_data;
1923    
1924  /* 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
1925  functions. */  functions. */
1926    
1927  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1928    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
1929    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
1930    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
1931    const uschar *start_used_ptr; /* Earliest consulted character */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
1932    const uschar *tables;         /* Character tables */    const pcre_uint8 *tables;         /* Character tables */
1933    int   start_offset;           /* The start offset value */    int   start_offset;               /* The start offset value */
1934    int   moptions;               /* Match options */    int   moptions;                   /* Match options */
1935    int   poptions;               /* Pattern options */    int   poptions;                   /* Pattern options */
1936    int    nltype;                /* Newline type */    int   nltype;                     /* Newline type */
1937    int    nllen;                 /* Newline string length */    int   nllen;                      /* Newline string length */
1938    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];                 /* Newline string when fixed */
1939    void  *callout_data;          /* To pass back to callouts */    void *callout_data;               /* To pass back to callouts */
1940      dfa_recursion_info *recursive;    /* Linked list of recursion data */
1941  } dfa_match_data;  } dfa_match_data;
1942    
1943  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1715  total length. */ Line 1973  total length. */
1973  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1974  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1975    
1976    /* Internal function prefix */
1977    
1978    #ifdef COMPILE_PCRE8
1979    #define PRIV(name) _pcre_##name
1980    #else
1981    #ifdef COMPILE_PCRE16
1982    #define PRIV(name) _pcre16_##name
1983    #else
1984    #error Unsupported compiling mode
1985    #endif /* COMPILE_PCRE16 */
1986    #endif /* COMPILE_PCRE8 */
1987    
1988  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
1989  codes. Each entry used to point directly to a name, but to reduce the number of  codes. Each entry used to point directly to a name, but to reduce the number of
1990  relocations in shared libraries, it now has an offset into a single string  relocations in shared libraries, it now has an offset into a single string
# Line 1732  of the exported public functions. They h Line 2002  of the exported public functions. They h
2002  but are not part of the PCRE public API. The data for these tables is in the  but are not part of the PCRE public API. The data for these tables is in the
2003  pcre_tables.c module. */  pcre_tables.c module. */
2004    
2005  extern const int    _pcre_utf8_table1[];  extern const int            PRIV(utf8_table1)[];
2006  extern const int    _pcre_utf8_table2[];  extern const int            PRIV(utf8_table2)[];
2007  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table3)[];
2008  extern const uschar _pcre_utf8_table4[];  extern const pcre_uint8     PRIV(utf8_table4)[];
2009    
2010  extern const int    _pcre_utf8_table1_size;  extern const int            PRIV(utf8_table1_size);
2011    
2012  extern const char   _pcre_utt_names[];  extern const char           PRIV(utt_names)[];
2013  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table PRIV(utt)[];
2014  extern const int _pcre_utt_size;  extern const int            PRIV(utt_size);
2015    
2016  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2017    
2018  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2019    
2020    
2021  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2022  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
2023  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2024    
2025  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  /* String comparison functions. */
2026  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);  #ifdef COMPILE_PCRE8
2027  extern int           _pcre_ord2utf8(int, uschar *);  
2028  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2029                         const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2030  extern int           _pcre_valid_utf8(USPTR, int);  #define STRCMP_UC_C8(str1, str2) \
2031  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);    strcmp((char *)(str1), (str2))
2032  extern BOOL          _pcre_xclass(int, const uschar *);  #define STRNCMP_UC_UC(str1, str2, num) \
2033      strncmp((char *)(str1), (char *)(str2), (num))
2034    #define STRNCMP_UC_C8(str1, str2, num) \
2035      strncmp((char *)(str1), (str2), (num))
2036    #define STRLEN_UC(str) strlen(str)
2037    
2038    #else
2039    
2040    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2041                               const pcre_uchar *);
2042    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2043                               const char *);
2044    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2045                               const pcre_uchar *, unsigned int num);
2046    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2047                               const char *, unsigned int num);
2048    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2049    
2050    #define STRCMP_UC_UC(str1, str2) \
2051      PRIV(strcmp_uc_uc)((str1), (str2))
2052    #define STRCMP_UC_C8(str1, str2) \
2053      PRIV(strcmp_uc_c8)((str1), (str2))
2054    #define STRNCMP_UC_UC(str1, str2, num) \
2055      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2056    #define STRNCMP_UC_C8(str1, str2, num) \
2057      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2058    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2059    
2060    #endif /* COMPILE_PCRE8 */
2061    
2062    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2063    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2064                               int *, BOOL);
2065    extern int               PRIV(ord2utf8)(int, pcre_uint8 *);
2066    extern real_pcre        *PRIV(try_flipped)(const real_pcre *, real_pcre *,
2067                               const pcre_study_data *, pcre_study_data *);
2068    #ifndef COMPILE_PCRE16
2069    extern int               PRIV(valid_utf8)(PCRE_PUCHAR, int, int *);
2070    #else
2071    extern int               PRIV(valid_utf16)(PCRE_PUCHAR, int, int *);
2072    #endif
2073    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2074                               int *, BOOL);
2075    extern BOOL              PRIV(xclass)(int, const pcre_uchar *);
2076    
2077    #ifdef SUPPORT_JIT
2078    extern void              PRIV(jit_compile)(const real_pcre *, pcre_extra *);
2079    extern int               PRIV(jit_exec)(const real_pcre *, void *, PCRE_SPTR,
2080                               int, int, int, int, int *, int);
2081    extern void              PRIV(jit_free)(void *);
2082    #endif
2083    
2084  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
2085    
2086  typedef struct {  typedef struct {
2087    uschar script;    pcre_uint8 script;
2088    uschar chartype;    pcre_uint8 chartype;
2089    pcre_int32 other_case;    pcre_int32 other_case;
2090  } ucd_record;  } ucd_record;
2091    
2092  extern const ucd_record  _pcre_ucd_records[];  extern const ucd_record  PRIV(ucd_records)[];
2093  extern const uschar      _pcre_ucd_stage1[];  extern const pcre_uint8  PRIV(ucd_stage1)[];
2094  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint16 PRIV(ucd_stage2)[];
2095  extern const int         _pcre_ucp_gentype[];  extern const int         PRIV(ucp_gentype)[];
2096    #ifdef SUPPORT_JIT
2097    extern const int         PRIV(ucp_typerange)[];
2098    #endif
2099    
2100  /* UCD access macros */  /* UCD access macros */
2101    
2102  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
2103  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (PRIV(ucd_records) + \
2104          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2105          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2106    
2107  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2108  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2109  #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]  #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2110  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2111    
2112  #endif  #endif

Legend:
Removed from v.483  
changed lines
  Added in v.764

  ViewVC Help
Powered by ViewVC 1.1.5