/[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 335 by ph10, Sat Apr 12 14:36:14 2008 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) 1997-2008 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
48    
49  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
50    
51  #if 0  #if 0
52  #define DEBUG  #define PCRE_DEBUG
53    #endif
54    
55    /* 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". */
62    
63    #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
64    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
65    #endif
66    
67    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
68    "configure" script ensures this, but not everybody uses "configure". */
69    
70    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
71    #define SUPPORT_UTF8 1
72  #endif  #endif
73    
74  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
# Line 60  It turns out that the Mac Debugging.h he Line 80  It turns out that the Mac Debugging.h he
80  be absolutely sure we get our version. */  be absolutely sure we get our version. */
81    
82  #undef DPRINTF  #undef DPRINTF
83  #ifdef DEBUG  #ifdef PCRE_DEBUG
84  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
85  #else  #else
86  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
# Line 72  setjmp and stdarg are used is when NO_RE Line 92  setjmp and stdarg are used is when NO_RE
92    
93  #include <ctype.h>  #include <ctype.h>
94  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
95  #include <stddef.h>  #include <stddef.h>
96  #include <stdio.h>  #include <stdio.h>
97  #include <stdlib.h>  #include <stdlib.h>
# Line 132  PCRE_EXP_DATA_DEFN only if they are not Line 150  PCRE_EXP_DATA_DEFN only if they are not
150  #  endif  #  endif
151  #endif  #endif
152    
153  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* When compiling with the MSVC compiler, it is sometimes necessary to include
154    a "calling convention" before exported function names. (This is secondhand
155    information; I know nothing about MSVC myself). For example, something like
156    
157      void __cdecl function(....)
158    
159    might be needed. In order so make this easy, all the exported functions have
160    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
161    set, we ensure here that it has no effect. */
162    
163    #ifndef PCRE_CALL_CONVENTION
164    #define PCRE_CALL_CONVENTION
165    #endif
166    
167    /* 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;
178  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
179    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
180      typedef int pcre_int16;
181  #else  #else
182    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
183  #endif  #endif
184    
185  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
186    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
187      typedef int pcre_int32;
188  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
189    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
190      typedef long int pcre_int32;
191  #else  #else
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
196    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
198    arithmetic. Handle this by defining a macro for the appropriate type. If
199    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
200    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
204    #include <stdint.h>
205    #elif HAVE_INTTYPES_H
206    #include <inttypes.h>
207    #endif
208    
209    #if defined INT64_MAX || defined int64_t
210    #define INT64_OR_DOUBLE int64_t
211    #else
212    #define INT64_OR_DOUBLE double
213    #endif
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 184  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 198  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 217  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 293  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 329  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 346  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 353  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
 #define NEXTCHAR(p) p++;  
495  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
496  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
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  /* Advance a character pointer one byte in non-UTF-8 mode and by one character  /* These macros were originally written in the form of loops that used data
506  in UTF-8 mode. */  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  #define NEXTCHAR(p) \  /* Base macro to pick up the remaining bytes of a UTF-8 character, not
511    p++; \  advancing the pointer. */
512    if (utf8) { while((*p & 0xc0) == 0x80) p++; }  
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          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
561              ((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        gcss -= 6; \        c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
574        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
575              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
576          eptr += 5; \
577        } \        } \
578      }      }
579    
# Line 421  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 454  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);
637      { \  
638      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
639      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
640      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
641      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
642      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
643        { \    c = *eptr; \
644        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       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 474  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 485  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. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
666    the restrictions on partial matching have been lifted. It remains for backwards
667    compatibility. */
668    
669  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
670  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 499  are in a 16-bit flags word. */ Line 672  are in a 16-bit flags word. */
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  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 510  time, run time, or study time, respectiv Line 685  time, run time, or study time, respectiv
685  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
686                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
687    
688  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
689    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
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_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_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
697     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
698       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
699    
700  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
701    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
702     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
703     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
704       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 548  variable-length repeat, or a anything ot Line 727  variable-length repeat, or a anything ot
727  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
728  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
729    
730  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
731    environments where these macros are defined elsewhere. Unfortunately, there
732    is no way to do the same for the typedef. */
733    
734  typedef int BOOL;  typedef int BOOL;
735    
736    #ifndef FALSE
737  #define FALSE   0  #define FALSE   0
738  #define TRUE    1  #define TRUE    1
739    #endif
740    
741    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
742    character constants like '*' because the compiler would emit their EBCDIC code,
743    which is different from their ASCII/UTF-8 code. Instead we define macros for
744    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
745    is enabled. When UTF-8 support is not enabled, the definitions use character
746    literals. Both character and string versions of each character are needed, and
747    there are some longer strings as well.
748    
749    This means that, on EBCDIC platforms, the PCRE library can handle either
750    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
751    would need different lookups depending on whether PCRE_UTF8 was set or not.
752    This would make it impossible to use characters in switch/case statements,
753    which would reduce performance. For a theoretical use (which nobody has asked
754    for) in a minority area (EBCDIC platforms), this is not sensible. Any
755    application that did need both could compile two versions of the library, using
756    macros to give the functions distinct names. */
757    
758    #ifndef SUPPORT_UTF8
759    
760    /* UTF-8 support is not enabled; use the platform-dependent character literals
761    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
762    
763    #define CHAR_HT                     '\t'
764    #define CHAR_VT                     '\v'
765    #define CHAR_FF                     '\f'
766    #define CHAR_CR                     '\r'
767    #define CHAR_NL                     '\n'
768    #define CHAR_BS                     '\b'
769    #define CHAR_BEL                    '\a'
770    #ifdef EBCDIC
771    #define CHAR_ESC                    '\047'
772    #define CHAR_DEL                    '\007'
773    #else
774    #define CHAR_ESC                    '\033'
775    #define CHAR_DEL                    '\177'
776    #endif
777    
778    #define CHAR_SPACE                  ' '
779    #define CHAR_EXCLAMATION_MARK       '!'
780    #define CHAR_QUOTATION_MARK         '"'
781    #define CHAR_NUMBER_SIGN            '#'
782    #define CHAR_DOLLAR_SIGN            '$'
783    #define CHAR_PERCENT_SIGN           '%'
784    #define CHAR_AMPERSAND              '&'
785    #define CHAR_APOSTROPHE             '\''
786    #define CHAR_LEFT_PARENTHESIS       '('
787    #define CHAR_RIGHT_PARENTHESIS      ')'
788    #define CHAR_ASTERISK               '*'
789    #define CHAR_PLUS                   '+'
790    #define CHAR_COMMA                  ','
791    #define CHAR_MINUS                  '-'
792    #define CHAR_DOT                    '.'
793    #define CHAR_SLASH                  '/'
794    #define CHAR_0                      '0'
795    #define CHAR_1                      '1'
796    #define CHAR_2                      '2'
797    #define CHAR_3                      '3'
798    #define CHAR_4                      '4'
799    #define CHAR_5                      '5'
800    #define CHAR_6                      '6'
801    #define CHAR_7                      '7'
802    #define CHAR_8                      '8'
803    #define CHAR_9                      '9'
804    #define CHAR_COLON                  ':'
805    #define CHAR_SEMICOLON              ';'
806    #define CHAR_LESS_THAN_SIGN         '<'
807    #define CHAR_EQUALS_SIGN            '='
808    #define CHAR_GREATER_THAN_SIGN      '>'
809    #define CHAR_QUESTION_MARK          '?'
810    #define CHAR_COMMERCIAL_AT          '@'
811    #define CHAR_A                      'A'
812    #define CHAR_B                      'B'
813    #define CHAR_C                      'C'
814    #define CHAR_D                      'D'
815    #define CHAR_E                      'E'
816    #define CHAR_F                      'F'
817    #define CHAR_G                      'G'
818    #define CHAR_H                      'H'
819    #define CHAR_I                      'I'
820    #define CHAR_J                      'J'
821    #define CHAR_K                      'K'
822    #define CHAR_L                      'L'
823    #define CHAR_M                      'M'
824    #define CHAR_N                      'N'
825    #define CHAR_O                      'O'
826    #define CHAR_P                      'P'
827    #define CHAR_Q                      'Q'
828    #define CHAR_R                      'R'
829    #define CHAR_S                      'S'
830    #define CHAR_T                      'T'
831    #define CHAR_U                      'U'
832    #define CHAR_V                      'V'
833    #define CHAR_W                      'W'
834    #define CHAR_X                      'X'
835    #define CHAR_Y                      'Y'
836    #define CHAR_Z                      'Z'
837    #define CHAR_LEFT_SQUARE_BRACKET    '['
838    #define CHAR_BACKSLASH              '\\'
839    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
840    #define CHAR_CIRCUMFLEX_ACCENT      '^'
841    #define CHAR_UNDERSCORE             '_'
842    #define CHAR_GRAVE_ACCENT           '`'
843    #define CHAR_a                      'a'
844    #define CHAR_b                      'b'
845    #define CHAR_c                      'c'
846    #define CHAR_d                      'd'
847    #define CHAR_e                      'e'
848    #define CHAR_f                      'f'
849    #define CHAR_g                      'g'
850    #define CHAR_h                      'h'
851    #define CHAR_i                      'i'
852    #define CHAR_j                      'j'
853    #define CHAR_k                      'k'
854    #define CHAR_l                      'l'
855    #define CHAR_m                      'm'
856    #define CHAR_n                      'n'
857    #define CHAR_o                      'o'
858    #define CHAR_p                      'p'
859    #define CHAR_q                      'q'
860    #define CHAR_r                      'r'
861    #define CHAR_s                      's'
862    #define CHAR_t                      't'
863    #define CHAR_u                      'u'
864    #define CHAR_v                      'v'
865    #define CHAR_w                      'w'
866    #define CHAR_x                      'x'
867    #define CHAR_y                      'y'
868    #define CHAR_z                      'z'
869    #define CHAR_LEFT_CURLY_BRACKET     '{'
870    #define CHAR_VERTICAL_LINE          '|'
871    #define CHAR_RIGHT_CURLY_BRACKET    '}'
872    #define CHAR_TILDE                  '~'
873    
874    #define STR_HT                      "\t"
875    #define STR_VT                      "\v"
876    #define STR_FF                      "\f"
877    #define STR_CR                      "\r"
878    #define STR_NL                      "\n"
879    #define STR_BS                      "\b"
880    #define STR_BEL                     "\a"
881    #ifdef EBCDIC
882    #define STR_ESC                     "\047"
883    #define STR_DEL                     "\007"
884    #else
885    #define STR_ESC                     "\033"
886    #define STR_DEL                     "\177"
887    #endif
888    
889    #define STR_SPACE                   " "
890    #define STR_EXCLAMATION_MARK        "!"
891    #define STR_QUOTATION_MARK          "\""
892    #define STR_NUMBER_SIGN             "#"
893    #define STR_DOLLAR_SIGN             "$"
894    #define STR_PERCENT_SIGN            "%"
895    #define STR_AMPERSAND               "&"
896    #define STR_APOSTROPHE              "'"
897    #define STR_LEFT_PARENTHESIS        "("
898    #define STR_RIGHT_PARENTHESIS       ")"
899    #define STR_ASTERISK                "*"
900    #define STR_PLUS                    "+"
901    #define STR_COMMA                   ","
902    #define STR_MINUS                   "-"
903    #define STR_DOT                     "."
904    #define STR_SLASH                   "/"
905    #define STR_0                       "0"
906    #define STR_1                       "1"
907    #define STR_2                       "2"
908    #define STR_3                       "3"
909    #define STR_4                       "4"
910    #define STR_5                       "5"
911    #define STR_6                       "6"
912    #define STR_7                       "7"
913    #define STR_8                       "8"
914    #define STR_9                       "9"
915    #define STR_COLON                   ":"
916    #define STR_SEMICOLON               ";"
917    #define STR_LESS_THAN_SIGN          "<"
918    #define STR_EQUALS_SIGN             "="
919    #define STR_GREATER_THAN_SIGN       ">"
920    #define STR_QUESTION_MARK           "?"
921    #define STR_COMMERCIAL_AT           "@"
922    #define STR_A                       "A"
923    #define STR_B                       "B"
924    #define STR_C                       "C"
925    #define STR_D                       "D"
926    #define STR_E                       "E"
927    #define STR_F                       "F"
928    #define STR_G                       "G"
929    #define STR_H                       "H"
930    #define STR_I                       "I"
931    #define STR_J                       "J"
932    #define STR_K                       "K"
933    #define STR_L                       "L"
934    #define STR_M                       "M"
935    #define STR_N                       "N"
936    #define STR_O                       "O"
937    #define STR_P                       "P"
938    #define STR_Q                       "Q"
939    #define STR_R                       "R"
940    #define STR_S                       "S"
941    #define STR_T                       "T"
942    #define STR_U                       "U"
943    #define STR_V                       "V"
944    #define STR_W                       "W"
945    #define STR_X                       "X"
946    #define STR_Y                       "Y"
947    #define STR_Z                       "Z"
948    #define STR_LEFT_SQUARE_BRACKET     "["
949    #define STR_BACKSLASH               "\\"
950    #define STR_RIGHT_SQUARE_BRACKET    "]"
951    #define STR_CIRCUMFLEX_ACCENT       "^"
952    #define STR_UNDERSCORE              "_"
953    #define STR_GRAVE_ACCENT            "`"
954    #define STR_a                       "a"
955    #define STR_b                       "b"
956    #define STR_c                       "c"
957    #define STR_d                       "d"
958    #define STR_e                       "e"
959    #define STR_f                       "f"
960    #define STR_g                       "g"
961    #define STR_h                       "h"
962    #define STR_i                       "i"
963    #define STR_j                       "j"
964    #define STR_k                       "k"
965    #define STR_l                       "l"
966    #define STR_m                       "m"
967    #define STR_n                       "n"
968    #define STR_o                       "o"
969    #define STR_p                       "p"
970    #define STR_q                       "q"
971    #define STR_r                       "r"
972    #define STR_s                       "s"
973    #define STR_t                       "t"
974    #define STR_u                       "u"
975    #define STR_v                       "v"
976    #define STR_w                       "w"
977    #define STR_x                       "x"
978    #define STR_y                       "y"
979    #define STR_z                       "z"
980    #define STR_LEFT_CURLY_BRACKET      "{"
981    #define STR_VERTICAL_LINE           "|"
982    #define STR_RIGHT_CURLY_BRACKET     "}"
983    #define STR_TILDE                   "~"
984    
985    #define STRING_ACCEPT0              "ACCEPT\0"
986    #define STRING_COMMIT0              "COMMIT\0"
987    #define STRING_F0                   "F\0"
988    #define STRING_FAIL0                "FAIL\0"
989    #define STRING_MARK0                "MARK\0"
990    #define STRING_PRUNE0               "PRUNE\0"
991    #define STRING_SKIP0                "SKIP\0"
992    #define STRING_THEN                 "THEN"
993    
994    #define STRING_alpha0               "alpha\0"
995    #define STRING_lower0               "lower\0"
996    #define STRING_upper0               "upper\0"
997    #define STRING_alnum0               "alnum\0"
998    #define STRING_ascii0               "ascii\0"
999    #define STRING_blank0               "blank\0"
1000    #define STRING_cntrl0               "cntrl\0"
1001    #define STRING_digit0               "digit\0"
1002    #define STRING_graph0               "graph\0"
1003    #define STRING_print0               "print\0"
1004    #define STRING_punct0               "punct\0"
1005    #define STRING_space0               "space\0"
1006    #define STRING_word0                "word\0"
1007    #define STRING_xdigit               "xdigit"
1008    
1009    #define STRING_DEFINE               "DEFINE"
1010    
1011    #define STRING_CR_RIGHTPAR             "CR)"
1012    #define STRING_LF_RIGHTPAR             "LF)"
1013    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1014    #define STRING_ANY_RIGHTPAR            "ANY)"
1015    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1016    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1017    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1018    #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 */
1023    
1024    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1025    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1026    only. */
1027    
1028    #define CHAR_HT                     '\011'
1029    #define CHAR_VT                     '\013'
1030    #define CHAR_FF                     '\014'
1031    #define CHAR_CR                     '\015'
1032    #define CHAR_NL                     '\012'
1033    #define CHAR_BS                     '\010'
1034    #define CHAR_BEL                    '\007'
1035    #define CHAR_ESC                    '\033'
1036    #define CHAR_DEL                    '\177'
1037    
1038    #define CHAR_SPACE                  '\040'
1039    #define CHAR_EXCLAMATION_MARK       '\041'
1040    #define CHAR_QUOTATION_MARK         '\042'
1041    #define CHAR_NUMBER_SIGN            '\043'
1042    #define CHAR_DOLLAR_SIGN            '\044'
1043    #define CHAR_PERCENT_SIGN           '\045'
1044    #define CHAR_AMPERSAND              '\046'
1045    #define CHAR_APOSTROPHE             '\047'
1046    #define CHAR_LEFT_PARENTHESIS       '\050'
1047    #define CHAR_RIGHT_PARENTHESIS      '\051'
1048    #define CHAR_ASTERISK               '\052'
1049    #define CHAR_PLUS                   '\053'
1050    #define CHAR_COMMA                  '\054'
1051    #define CHAR_MINUS                  '\055'
1052    #define CHAR_DOT                    '\056'
1053    #define CHAR_SLASH                  '\057'
1054    #define CHAR_0                      '\060'
1055    #define CHAR_1                      '\061'
1056    #define CHAR_2                      '\062'
1057    #define CHAR_3                      '\063'
1058    #define CHAR_4                      '\064'
1059    #define CHAR_5                      '\065'
1060    #define CHAR_6                      '\066'
1061    #define CHAR_7                      '\067'
1062    #define CHAR_8                      '\070'
1063    #define CHAR_9                      '\071'
1064    #define CHAR_COLON                  '\072'
1065    #define CHAR_SEMICOLON              '\073'
1066    #define CHAR_LESS_THAN_SIGN         '\074'
1067    #define CHAR_EQUALS_SIGN            '\075'
1068    #define CHAR_GREATER_THAN_SIGN      '\076'
1069    #define CHAR_QUESTION_MARK          '\077'
1070    #define CHAR_COMMERCIAL_AT          '\100'
1071    #define CHAR_A                      '\101'
1072    #define CHAR_B                      '\102'
1073    #define CHAR_C                      '\103'
1074    #define CHAR_D                      '\104'
1075    #define CHAR_E                      '\105'
1076    #define CHAR_F                      '\106'
1077    #define CHAR_G                      '\107'
1078    #define CHAR_H                      '\110'
1079    #define CHAR_I                      '\111'
1080    #define CHAR_J                      '\112'
1081    #define CHAR_K                      '\113'
1082    #define CHAR_L                      '\114'
1083    #define CHAR_M                      '\115'
1084    #define CHAR_N                      '\116'
1085    #define CHAR_O                      '\117'
1086    #define CHAR_P                      '\120'
1087    #define CHAR_Q                      '\121'
1088    #define CHAR_R                      '\122'
1089    #define CHAR_S                      '\123'
1090    #define CHAR_T                      '\124'
1091    #define CHAR_U                      '\125'
1092    #define CHAR_V                      '\126'
1093    #define CHAR_W                      '\127'
1094    #define CHAR_X                      '\130'
1095    #define CHAR_Y                      '\131'
1096    #define CHAR_Z                      '\132'
1097    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1098    #define CHAR_BACKSLASH              '\134'
1099    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1100    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1101    #define CHAR_UNDERSCORE             '\137'
1102    #define CHAR_GRAVE_ACCENT           '\140'
1103    #define CHAR_a                      '\141'
1104    #define CHAR_b                      '\142'
1105    #define CHAR_c                      '\143'
1106    #define CHAR_d                      '\144'
1107    #define CHAR_e                      '\145'
1108    #define CHAR_f                      '\146'
1109    #define CHAR_g                      '\147'
1110    #define CHAR_h                      '\150'
1111    #define CHAR_i                      '\151'
1112    #define CHAR_j                      '\152'
1113    #define CHAR_k                      '\153'
1114    #define CHAR_l                      '\154'
1115    #define CHAR_m                      '\155'
1116    #define CHAR_n                      '\156'
1117    #define CHAR_o                      '\157'
1118    #define CHAR_p                      '\160'
1119    #define CHAR_q                      '\161'
1120    #define CHAR_r                      '\162'
1121    #define CHAR_s                      '\163'
1122    #define CHAR_t                      '\164'
1123    #define CHAR_u                      '\165'
1124    #define CHAR_v                      '\166'
1125    #define CHAR_w                      '\167'
1126    #define CHAR_x                      '\170'
1127    #define CHAR_y                      '\171'
1128    #define CHAR_z                      '\172'
1129    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1130    #define CHAR_VERTICAL_LINE          '\174'
1131    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1132    #define CHAR_TILDE                  '\176'
1133    
1134    #define STR_HT                      "\011"
1135    #define STR_VT                      "\013"
1136    #define STR_FF                      "\014"
1137    #define STR_CR                      "\015"
1138    #define STR_NL                      "\012"
1139    #define STR_BS                      "\010"
1140    #define STR_BEL                     "\007"
1141    #define STR_ESC                     "\033"
1142    #define STR_DEL                     "\177"
1143    
1144    #define STR_SPACE                   "\040"
1145    #define STR_EXCLAMATION_MARK        "\041"
1146    #define STR_QUOTATION_MARK          "\042"
1147    #define STR_NUMBER_SIGN             "\043"
1148    #define STR_DOLLAR_SIGN             "\044"
1149    #define STR_PERCENT_SIGN            "\045"
1150    #define STR_AMPERSAND               "\046"
1151    #define STR_APOSTROPHE              "\047"
1152    #define STR_LEFT_PARENTHESIS        "\050"
1153    #define STR_RIGHT_PARENTHESIS       "\051"
1154    #define STR_ASTERISK                "\052"
1155    #define STR_PLUS                    "\053"
1156    #define STR_COMMA                   "\054"
1157    #define STR_MINUS                   "\055"
1158    #define STR_DOT                     "\056"
1159    #define STR_SLASH                   "\057"
1160    #define STR_0                       "\060"
1161    #define STR_1                       "\061"
1162    #define STR_2                       "\062"
1163    #define STR_3                       "\063"
1164    #define STR_4                       "\064"
1165    #define STR_5                       "\065"
1166    #define STR_6                       "\066"
1167    #define STR_7                       "\067"
1168    #define STR_8                       "\070"
1169    #define STR_9                       "\071"
1170    #define STR_COLON                   "\072"
1171    #define STR_SEMICOLON               "\073"
1172    #define STR_LESS_THAN_SIGN          "\074"
1173    #define STR_EQUALS_SIGN             "\075"
1174    #define STR_GREATER_THAN_SIGN       "\076"
1175    #define STR_QUESTION_MARK           "\077"
1176    #define STR_COMMERCIAL_AT           "\100"
1177    #define STR_A                       "\101"
1178    #define STR_B                       "\102"
1179    #define STR_C                       "\103"
1180    #define STR_D                       "\104"
1181    #define STR_E                       "\105"
1182    #define STR_F                       "\106"
1183    #define STR_G                       "\107"
1184    #define STR_H                       "\110"
1185    #define STR_I                       "\111"
1186    #define STR_J                       "\112"
1187    #define STR_K                       "\113"
1188    #define STR_L                       "\114"
1189    #define STR_M                       "\115"
1190    #define STR_N                       "\116"
1191    #define STR_O                       "\117"
1192    #define STR_P                       "\120"
1193    #define STR_Q                       "\121"
1194    #define STR_R                       "\122"
1195    #define STR_S                       "\123"
1196    #define STR_T                       "\124"
1197    #define STR_U                       "\125"
1198    #define STR_V                       "\126"
1199    #define STR_W                       "\127"
1200    #define STR_X                       "\130"
1201    #define STR_Y                       "\131"
1202    #define STR_Z                       "\132"
1203    #define STR_LEFT_SQUARE_BRACKET     "\133"
1204    #define STR_BACKSLASH               "\134"
1205    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1206    #define STR_CIRCUMFLEX_ACCENT       "\136"
1207    #define STR_UNDERSCORE              "\137"
1208    #define STR_GRAVE_ACCENT            "\140"
1209    #define STR_a                       "\141"
1210    #define STR_b                       "\142"
1211    #define STR_c                       "\143"
1212    #define STR_d                       "\144"
1213    #define STR_e                       "\145"
1214    #define STR_f                       "\146"
1215    #define STR_g                       "\147"
1216    #define STR_h                       "\150"
1217    #define STR_i                       "\151"
1218    #define STR_j                       "\152"
1219    #define STR_k                       "\153"
1220    #define STR_l                       "\154"
1221    #define STR_m                       "\155"
1222    #define STR_n                       "\156"
1223    #define STR_o                       "\157"
1224    #define STR_p                       "\160"
1225    #define STR_q                       "\161"
1226    #define STR_r                       "\162"
1227    #define STR_s                       "\163"
1228    #define STR_t                       "\164"
1229    #define STR_u                       "\165"
1230    #define STR_v                       "\166"
1231    #define STR_w                       "\167"
1232    #define STR_x                       "\170"
1233    #define STR_y                       "\171"
1234    #define STR_z                       "\172"
1235    #define STR_LEFT_CURLY_BRACKET      "\173"
1236    #define STR_VERTICAL_LINE           "\174"
1237    #define STR_RIGHT_CURLY_BRACKET     "\175"
1238    #define STR_TILDE                   "\176"
1239    
1240    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1241    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1242    #define STRING_F0                   STR_F "\0"
1243    #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"
1246    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1247    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1248    
1249    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1250    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1251    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1252    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1253    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1254    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1255    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1256    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1257    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1258    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1259    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1260    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1261    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1262    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1263    
1264    #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
1267    #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
1269    #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
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
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
1273    #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 */
1278    
1279  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1280    
1281  #ifndef ESC_e  #ifndef ESC_e
1282  #define ESC_e 27  #define ESC_e CHAR_ESC
1283  #endif  #endif
1284    
1285  #ifndef ESC_f  #ifndef ESC_f
1286  #define ESC_f '\f'  #define ESC_f CHAR_FF
1287  #endif  #endif
1288    
1289  #ifndef ESC_n  #ifndef ESC_n
1290  #define ESC_n '\n'  #define ESC_n CHAR_NL
1291  #endif  #endif
1292    
1293  #ifndef ESC_r  #ifndef ESC_r
1294  #define ESC_r '\r'  #define ESC_r CHAR_CR
1295  #endif  #endif
1296    
1297  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1298  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1299    
1300  #ifndef ESC_tee  #ifndef ESC_tee
1301  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1302  #endif  #endif
1303    
1304  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
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 603  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. The final one must be  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1333  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1334  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1335  ESC_Z to detect the types that may be repeated. These are the types that  
1336  consume characters. If any new escapes are put in between that don't consume a  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1337  character, that code will have to change. */  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
1342    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1343    greater than ESC_b and less than ESC_Z to detect the types that may be
1344    repeated. These are the types that consume characters. If any new escapes are
1345    put in between that don't consume a character, that code will have to change.
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_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1350         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, ESC_REF };         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 };
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    
1358  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1359  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There are also tables called
1360  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1361    
1362  enum {  enum {
1363    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 639  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 */
   OP_ANY,            /* 12 Match any character */  
   OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */  
   OP_NOTPROP,        /* 14 \P (not Unicode property) */  
   OP_PROP,           /* 15 \p (Unicode property) */  
   OP_ANYNL,          /* 16 \R (any newline sequence) */  
   OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */  
   OP_HSPACE,         /* 18 \h (horizontal whitespace) */  
   OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */  
   OP_VSPACE,         /* 20 \v (vertical whitespace) */  
   OP_EXTUNI,         /* 21 \X (extended Unicode sequence */  
   OP_EODN,           /* 22 End of data or \n at end of data: \Z. */  
   OP_EOD,            /* 23 End of data: \z */  
   
   OP_OPT,            /* 24 Set runtime options */  
   OP_CIRC,           /* 25 Start of line - varies with multiline switch */  
   OP_DOLL,           /* 26 End of line - varies with multiline switch */  
   OP_CHAR,           /* 27 Match one character, casefully */  
   OP_CHARNC,         /* 28 Match one character, caselessly */  
   OP_NOT,            /* 29 Match one character, not the following one */  
   
   OP_STAR,           /* 30 The maximizing and minimizing versions of */  
   OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */  
   OP_PLUS,           /* 32 the minimizing one second. */  
   OP_MINPLUS,        /* 33 This first set applies to single characters.*/  
   OP_QUERY,          /* 34 */  
   OP_MINQUERY,       /* 35 */  
   
   OP_UPTO,           /* 36 From 0 to n matches */  
   OP_MINUPTO,        /* 37 */  
   OP_EXACT,          /* 38 Exactly n matches */  
   
   OP_POSSTAR,        /* 39 Possessified star */  
   OP_POSPLUS,        /* 40 Possessified plus */  
   OP_POSQUERY,       /* 41 Posesssified query */  
   OP_POSUPTO,        /* 42 Possessified upto */  
   
   OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
   OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */  
   OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */  
   OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */  
   OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */  
   OP_NOTMINQUERY,    /* 48 */  
   
   OP_NOTUPTO,        /* 49 From 0 to n matches */  
   OP_NOTMINUPTO,     /* 50 */  
   OP_NOTEXACT,       /* 51 Exactly n matches */  
   
   OP_NOTPOSSTAR,     /* 52 Possessified versions */  
   OP_NOTPOSPLUS,     /* 53 */  
   OP_NOTPOSQUERY,    /* 54 */  
   OP_NOTPOSUPTO,     /* 55 */  
   
   OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
   OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */  
   OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */  
   OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */  
   OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */  
   OP_TYPEMINQUERY,   /* 61 */  
   
   OP_TYPEUPTO,       /* 62 From 0 to n matches */  
   OP_TYPEMINUPTO,    /* 63 */  
   OP_TYPEEXACT,      /* 64 Exactly n matches */  
   
   OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
   OP_TYPEPOSPLUS,    /* 66 */  
   OP_TYPEPOSQUERY,   /* 67 */  
   OP_TYPEPOSUPTO,    /* 68 */  
   
   OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
   OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */  
   OP_CRPLUS,         /* 71 the minimizing one second. These codes must */  
   OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */  
   OP_CRQUERY,        /* 73 These are for character classes and back refs */  
   OP_CRMINQUERY,     /* 74 */  
   OP_CRRANGE,        /* 75 These are different to the three sets above. */  
   OP_CRMINRANGE,     /* 76 */  
   
   OP_CLASS,          /* 77 Match a character class, chars < 256 only */  
   OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative  
                            class - the difference is relevant only when a UTF-8  
                            character > 255 is encountered. */  
   
   OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the  
                            class. This does both positive and negative. */  
   
   OP_REF,            /* 80 Match a back reference */  
   OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */  
   OP_CALLOUT,        /* 82 Call out to external function if provided */  
   
   OP_ALT,            /* 83 Start of alternation */  
   OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */  
   OP_KETRMAX,        /* 85 These two must remain together and in this */  
   OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */  
   
   /* The assertions must come before BRA, CBRA, ONCE, and COND.*/  
   
   OP_ASSERT,         /* 87 Positive lookahead */  
   OP_ASSERT_NOT,     /* 88 Negative lookahead */  
   OP_ASSERTBACK,     /* 89 Positive lookbehind */  
   OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */  
   OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */  
   
   /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,  
   as there's a test for >= ONCE for a subpattern that isn't an assertion. */  
   
   OP_ONCE,           /* 92 Atomic group */  
   OP_BRA,            /* 93 Start of non-capturing bracket */  
   OP_CBRA,           /* 94 Start of capturing bracket */  
   OP_COND,           /* 95 Conditional group */  
1378    
1379    /* These three must follow the previous three, in the same order. There's a    OP_ANY,            /* 12 Match any character except newline */
1380    check for >= SBRA to distinguish the two sets. */    OP_ALLANY,         /* 13 Match any character */
1381      OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1382      OP_NOTPROP,        /* 15 \P (not Unicode property) */
1383      OP_PROP,           /* 16 \p (Unicode property) */
1384      OP_ANYNL,          /* 17 \R (any newline sequence) */
1385      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1386      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1387      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1388      OP_VSPACE,         /* 21 \v (vertical whitespace) */
1389      OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1390      OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1391      OP_EOD,            /* 24 End of data: \z */
1392    
1393      OP_CIRC,           /* 25 Start of line - not multiline */
1394      OP_CIRCM,          /* 26 Start of line - multiline */
1395      OP_DOLL,           /* 27 End of line - not multiline */
1396      OP_DOLLM,          /* 28 End of line - multiline */
1397      OP_CHAR,           /* 29 Match one character, casefully */
1398      OP_CHARI,          /* 30 Match one character, caselessly */
1399      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1400      OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1401    
1402      /* The following sets of 13 opcodes must always be kept in step because
1403      the offset from the first one is used to generate the others. */
1404    
1405      /**** Single characters, caseful, must precede the caseless ones ****/
1406    
1407      OP_STAR,           /* 33 The maximizing and minimizing versions of */
1408      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1409      OP_PLUS,           /* 35 the minimizing one second. */
1410      OP_MINPLUS,        /* 36 */
1411      OP_QUERY,          /* 37 */
1412      OP_MINQUERY,       /* 38 */
1413    
1414      OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1415      OP_MINUPTO,        /* 40 */
1416      OP_EXACT,          /* 41 Exactly n matches */
1417    
1418      OP_POSSTAR,        /* 42 Possessified star, caseful */
1419      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1420      OP_POSQUERY,       /* 44 Posesssified query, caseful */
1421      OP_POSUPTO,        /* 45 Possessified upto, caseful */
1422    
1423      /**** Single characters, caseless, must follow the caseful ones */
1424    
1425      OP_STARI,          /* 46 */
1426      OP_MINSTARI,       /* 47 */
1427      OP_PLUSI,          /* 48 */
1428      OP_MINPLUSI,       /* 49 */
1429      OP_QUERYI,         /* 50 */
1430      OP_MINQUERYI,      /* 51 */
1431    
1432      OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1433      OP_MINUPTOI,       /* 53 */
1434      OP_EXACTI,         /* 54 */
1435    
1436      OP_POSSTARI,       /* 55 Possessified star, caseless */
1437      OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1438      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1439      OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1440    
1441      /**** The negated ones must follow the non-negated ones, and match them ****/
1442      /**** Negated single character, caseful; must precede the caseless ones ****/
1443    
1444      OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1445      OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1446      OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1447      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1448      OP_NOTQUERY,       /* 63 */
1449      OP_NOTMINQUERY,    /* 64 */
1450    
1451      OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1452      OP_NOTMINUPTO,     /* 66 */
1453      OP_NOTEXACT,       /* 67 Exactly n matches */
1454    
1455      OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1456      OP_NOTPOSPLUS,     /* 69 */
1457      OP_NOTPOSQUERY,    /* 70 */
1458      OP_NOTPOSUPTO,     /* 71 */
1459    
1460      /**** Negated single character, caseless; must follow the caseful ones ****/
1461    
1462      OP_NOTSTARI,       /* 72 */
1463      OP_NOTMINSTARI,    /* 73 */
1464      OP_NOTPLUSI,       /* 74 */
1465      OP_NOTMINPLUSI,    /* 75 */
1466      OP_NOTQUERYI,      /* 76 */
1467      OP_NOTMINQUERYI,   /* 77 */
1468    
1469      OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1470      OP_NOTMINUPTOI,    /* 79 */
1471      OP_NOTEXACTI,      /* 80 Exactly n matches */
1472    
1473      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1474      OP_NOTPOSPLUSI,    /* 82 */
1475      OP_NOTPOSQUERYI,   /* 83 */
1476      OP_NOTPOSUPTOI,    /* 84 */
1477    
1478      /**** Character types ****/
1479    
1480      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1481      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1482      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1483      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1484      OP_TYPEQUERY,      /* 89 */
1485      OP_TYPEMINQUERY,   /* 90 */
1486    
1487      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1488      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    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    /* These five must follow the previous five, in the same order. There's a
1551    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    check for >= SBRA to distinguish the two sets. */
   OP_SCOND,          /* 98 Conditional group, check empty */  
   
   OP_CREF,           /* 99 Used to hold a capture number as condition */  
   OP_RREF,           /* 100 Used to hold a recursion number as condition */  
   OP_DEF,            /* 101 The DEFINE condition */  
1552    
1553    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1554    OP_BRAMINZERO,     /* 103 order. */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1555      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. */
1560    
1561      OP_CREF,           /* 135 Used to hold a capture number as condition */
1562      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1563      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1564      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1565      OP_DEF,            /* 139 The DEFINE condition */
1566    
1567      OP_BRAZERO,        /* 140 These two must remain together and in this */
1568      OP_BRAMINZERO,     /* 141 order. */
1569      OP_BRAPOSZERO,     /* 142 */
1570    
1571    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1572    
1573    OP_PRUNE,          /* 104 */    OP_MARK,           /* 143 always has an argument */
1574    OP_SKIP,           /* 105 */    OP_PRUNE,          /* 144 */
1575    OP_THEN,           /* 106 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1576    OP_COMMIT,         /* 107 */    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,           /* 108 */    OP_FAIL,           /* 151 */
1585    OP_ACCEPT,         /* 109 */    OP_ACCEPT,         /* 152 */
1586      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        /* 110 */    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
1601    definitions that follow must also be updated to match. There are also tables
1602    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", "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    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once", "Once_NC",                                              \
1636    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Bra", "BraPos", "CBra", "CBraPos",                             \
1637    "Skip zero"    "Cond",                                                         \
1638      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1639      "SCond",                                                        \
1640      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1641      "Brazero", "Braminzero", "Braposzero",                          \
1642      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1643      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1644      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1645      "Close", "Skip zero"
1646    
1647    
1648  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 819  in UTF-8 mode. The code that uses this t Line 1658  in UTF-8 mode. The code that uses this t
1658    1,                             /* End                                    */ \    1,                             /* End                                    */ \
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,                          /* Any, 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 845  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,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1721    3,                             /* RREF                                   */ \    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, 1                        /* FAIL, ACCEPT, 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"
1731  /* A magic value for OP_RREF to indicate the "any recursion" 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 884  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 };         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 900  Because people can now save and re-use c Line 1759  Because people can now save and re-use c
1759  structure should be made at the end, and something earlier (e.g. a new  structure should be made at the end, and something earlier (e.g. a new
1760  flag in the options or one of the dummy fields) should indicate that the new  flag in the options or one of the dummy fields) should indicate that the new
1761  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1762  NOTE NOTE NOTE:  NOTE NOTE NOTE
1763  */  */
1764    
1765  typedef struct real_pcre {  typedef struct real_pcre {
# Line 918  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 927  remark (see NOTE above) about extending Line 1786  remark (see NOTE above) about extending
1786    
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 options;    pcre_uint32 flags;              /* Private flags */
1790    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
1791      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
1795    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1796    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 {
1800      struct open_capitem *next;    /* Chain link */
1801      pcre_uint16 number;           /* Capture number */
1802      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1803    } 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    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
1818    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
1819    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
1820    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
1821    int  bracount;                /* Count of capturing parens as we compile */    int  name_entry_size;             /* Size of each entry */
1822    int  final_bracount;          /* Saved value after first pass */    int  bracount;                    /* Count of capturing parens as we compile */
1823    int  top_backref;             /* Maximum back reference */    int  final_bracount;              /* Saved value after first pass */
1824    unsigned int backref_map;     /* Bitmap of low back refs */    int  top_backref;                 /* Maximum back reference */
1825    int  external_options;        /* External (initial) options */    unsigned int backref_map;         /* Bitmap of low back refs */
1826    int  external_flags;          /* External flag bits to be set */    int  assert_depth;                /* Depth of nested assertions */
1827    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_options;            /* External (initial) options */
1828    BOOL had_accept;              /* (*ACCEPT) encountered */    int  external_flags;              /* External flag bits to be set */
1829    int  nltype;                  /* Newline type */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1830    int  nllen;                   /* Newline string length */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1831    uschar nl[4];                 /* Newline string when fixed length */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1832      int  nltype;                      /* Newline type */
1833      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;    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 */  
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 997  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    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
1888    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
1889    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
1890    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
1891    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
1892    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
1893    BOOL   utf8;                  /* UTF8 flag */    BOOL   offset_overflow;         /* Set if too many extractions */
1894    BOOL   endonly;               /* Dollar not before final \n */    BOOL   notbol;                  /* NOTBOL flag */
1895    BOOL   notempty;              /* Empty string match not wanted */    BOOL   noteol;                  /* NOTEOL flag */
1896    BOOL   partial;               /* PARTIAL flag */    BOOL   utf8;                    /* UTF8 flag */
1897    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
1898    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   use_ucp;                 /* PCRE_UCP flag */
1899    const uschar *start_code;     /* For use when recursing */    BOOL   endonly;                 /* Dollar not before final \n */
1900    USPTR  start_subject;         /* Start of the subject string */    BOOL   notempty;                /* Empty string match not wanted */
1901    USPTR  end_subject;           /* End of the subject string */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
1902    USPTR  start_match_ptr;       /* Start of matched string */    BOOL   hitend;                  /* Hit the end of the subject at some point */
1903    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
1904    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
1905    int    capture_last;          /* Most recent capture number */    const  pcre_uchar *start_code;  /* For use when recursing */
1906    int    start_offset;          /* The start offset value */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
1907    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    PCRE_PUCHAR end_subject;        /* End of the subject string */
1908    int    eptrn;                 /* Next free eptrblock */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
1909    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
1910    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
1911      int    partial;                 /* PARTIAL options */
1912      int    end_offset_top;          /* Highwater mark at end of match */
1913      int    capture_last;            /* Most recent capture number */
1914      int    start_offset;            /* The start offset value */
1915      int    match_function_type;     /* Set for certain special calls of MATCH() */
1916      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 *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
1932    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
1933    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
1934    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
1935    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
1936    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
1937    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
1938      pcre_uchar nl[4];                 /* Newline string when fixed */
1939      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 1077  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 1094  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 BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  /* String comparison functions. */
2026                        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_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
2031  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
2032  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
2033  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,    strncmp((char *)(str1), (char *)(str2), (num))
2034                        int *, BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
2035  extern BOOL         _pcre_xclass(int, const uschar *);    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) */
2085    
2086    typedef struct {
2087      pcre_uint8 script;
2088      pcre_uint8 chartype;
2089      pcre_int32 other_case;
2090    } ucd_record;
2091    
2092    extern const ucd_record  PRIV(ucd_records)[];
2093    extern const pcre_uint8  PRIV(ucd_stage1)[];
2094    extern const pcre_uint16 PRIV(ucd_stage2)[];
2095    extern const int         PRIV(ucp_gentype)[];
2096    #ifdef SUPPORT_JIT
2097    extern const int         PRIV(ucp_typerange)[];
2098    #endif
2099    
2100    /* UCD access macros */
2101    
2102    #define UCD_BLOCK_SIZE 128
2103    #define GET_UCD(ch) (PRIV(ucd_records) + \
2104            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2105            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2106    
2107    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2108    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2109    #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2110    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2111    
2112  #endif  #endif
2113    

Legend:
Removed from v.335  
changed lines
  Added in v.764

  ViewVC Help
Powered by ViewVC 1.1.5