/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Diff of /code/trunk/pcre_internal.h

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

revision 412 by ph10, Sat Apr 11 10:34:37 2009 UTC revision 1044 by ph10, Thu Sep 20 16:23:57 2012 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2009 University of Cambridge             Copyright (c) 1997-2012 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  #endif
54    
55  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"  /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56  script prevents both being selected, but not everybody uses "configure". */  #ifndef COMPILE_PCRE16
57    #define COMPILE_PCRE8
 #if defined EBCDIC && defined SUPPORT_UTF8  
 #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.  
58  #endif  #endif
59    
60  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The  /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
61  "configure" script ensures this, but not everybody uses "configure". */  "configure" script ensures this, but not everybody uses "configure". */
62    
63  #if defined SUPPORT_UCP && !defined SUPPORT_UTF8  #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
64    #define SUPPORT_UTF 1
65    #endif
66    
67    /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
68    reasons with existing code. */
69    
70    #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
71    #define SUPPORT_UTF 1
72    #endif
73    
74    /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
75    Until then we define it if SUPPORT_UTF is defined. */
76    
77    #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
78  #define SUPPORT_UTF8 1  #define SUPPORT_UTF8 1
79  #endif  #endif
80    
81    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
82    script prevents both being selected, but not everybody uses "configure". */
83    
84    #if defined EBCDIC && defined SUPPORT_UTF
85    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
86    #endif
87    
88  /* 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
89  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
90  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
# Line 74  It turns out that the Mac Debugging.h he Line 94  It turns out that the Mac Debugging.h he
94  be absolutely sure we get our version. */  be absolutely sure we get our version. */
95    
96  #undef DPRINTF  #undef DPRINTF
97  #ifdef DEBUG  #ifdef PCRE_DEBUG
98  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
99  #else  #else
100  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
# Line 86  setjmp and stdarg are used is when NO_RE Line 106  setjmp and stdarg are used is when NO_RE
106    
107  #include <ctype.h>  #include <ctype.h>
108  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
109  #include <stddef.h>  #include <stddef.h>
110  #include <stdio.h>  #include <stdio.h>
111  #include <stdlib.h>  #include <stdlib.h>
# Line 160  set, we ensure here that it has no effec Line 178  set, we ensure here that it has no effec
178  #define PCRE_CALL_CONVENTION  #define PCRE_CALL_CONVENTION
179  #endif  #endif
180    
181  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
182  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
183  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
184  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
185  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
186    
187    typedef unsigned char pcre_uint8;
188    
189  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
190    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
191    typedef short pcre_int16;    typedef short pcre_int16;
# Line 186  preprocessor time in standard C environm Line 206  preprocessor time in standard C environm
206    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
207  #endif  #endif
208    
209    /* When checking for integer overflow in pcre_compile(), we need to handle
210    large integers. If a 64-bit integer type is available, we can use that.
211    Otherwise we have to cast to double, which of course requires floating point
212    arithmetic. Handle this by defining a macro for the appropriate type. If
213    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
214    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
215    by "configure". */
216    
217    #if HAVE_STDINT_H
218    #include <stdint.h>
219    #elif HAVE_INTTYPES_H
220    #include <inttypes.h>
221    #endif
222    
223    #if defined INT64_MAX || defined int64_t
224    #define INT64_OR_DOUBLE int64_t
225    #else
226    #define INT64_OR_DOUBLE double
227    #endif
228    
229  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
230  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
231  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 *,
232  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. This character type is
233  to save lots of typing. I tried "uchar", but it causes problems on Digital  called pcre_uchar.
234  Unix, where it is defined in sys/types, so use "uschar" instead. */  
235    The IN_UCHARS macro multiply its argument with the byte size of the current
236    pcre_uchar type. Useful for memcpy and such operations, whose require the
237    byte size of their input/output buffers.
238    
239    The MAX_255 macro checks whether its pcre_uchar input is less than 256.
240    
241    The TABLE_GET macro is designed for accessing elements of tables whose contain
242    exactly 256 items. When the character is able to contain more than 256
243    items, some check is needed before accessing these tables.
244    */
245    
246    #ifdef COMPILE_PCRE8
247    
248    typedef unsigned char pcre_uchar;
249    #define IN_UCHARS(x) (x)
250    #define MAX_255(c) 1
251    #define TABLE_GET(c, table, default) ((table)[c])
252    
253    #else
254    
255    #ifdef COMPILE_PCRE16
256    #if USHRT_MAX != 65535
257    /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
258    pcre.h(.in) and disable (comment out) this message. */
259    #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
260    #endif
261    
262    typedef pcre_uint16 pcre_uchar;
263    #define IN_UCHARS(x) ((x) << 1)
264    #define MAX_255(c) ((c) <= 255u)
265    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
266    
267    #else
268    #error Unsupported compiling mode
269    #endif /* COMPILE_PCRE16 */
270    
271  typedef unsigned char uschar;  #endif /* COMPILE_PCRE8 */
272    
273  /* 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
274  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 216  start/end of string field names are. */ Line 291  start/end of string field names are. */
291  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
292    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
293      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
294       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
295         utf8)) \         &(NLBLOCK->nllen), utf)) \
296      : \      : \
297      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
298       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 230  start/end of string field names are. */ Line 305  start/end of string field names are. */
305  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
306    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
307      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
308       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
309         &(NLBLOCK->nllen), utf8)) \         &(NLBLOCK->nllen), utf)) \
310      : \      : \
311      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
312       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 249  used for the external interface and appe Line 324  used for the external interface and appe
324  must begin with PCRE_. */  must begin with PCRE_. */
325    
326  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
327  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
328  #else  #else
329  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
330  #endif  #endif
331    
   
   
332  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
333  values. */  values. */
334    
# Line 325  The macros are controlled by the value o Line 396  The macros are controlled by the value o
396  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
397  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
398    
399    #ifdef COMPILE_PCRE8
400    
401  #if LINK_SIZE == 2  #if LINK_SIZE == 2
402    
403  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 361  is automated on Unix systems via the "co Line 434  is automated on Unix systems via the "co
434  #define GET(a,n) \  #define GET(a,n) \
435    (((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])
436    
437  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
438    #define MAX_PATTERN_SIZE (1 << 30)
439    
440    #else
441    #error LINK_SIZE must be either 2, 3, or 4
442    #endif
443    
444    #else /* COMPILE_PCRE8 */
445    
446    #ifdef COMPILE_PCRE16
447    
448    #if LINK_SIZE == 2
449    
450    #undef LINK_SIZE
451    #define LINK_SIZE 1
452    
453    #define PUT(a,n,d)   \
454      (a[n] = (d))
455    
456    #define GET(a,n) \
457      (a[n])
458    
459    #define MAX_PATTERN_SIZE (1 << 16)
460    
461    #elif LINK_SIZE == 3 || LINK_SIZE == 4
462    
463    #undef LINK_SIZE
464    #define LINK_SIZE 2
465    
466    #define PUT(a,n,d)   \
467      (a[n] = (d) >> 16), \
468      (a[(n)+1] = (d) & 65535)
469    
470    #define GET(a,n) \
471      (((a)[n] << 16) | (a)[(n)+1])
472    
473    /* Keep it positive */
474    #define MAX_PATTERN_SIZE (1 << 30)
475    
476  #else  #else
477  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
478  #endif  #endif
479    
480    #else
481    #error Unsupported compiling mode
482    #endif /* COMPILE_PCRE16 */
483    
484    #endif /* COMPILE_PCRE8 */
485    
486  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
487    
# Line 378  is automated on Unix systems via the "co Line 492  is automated on Unix systems via the "co
492  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
493  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
494    
495    #ifdef COMPILE_PCRE8
496    
497    #define IMM2_SIZE 2
498    
499  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
500    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
501    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 385  capturing parenthesis numbers in back re Line 503  capturing parenthesis numbers in back re
503  #define GET2(a,n) \  #define GET2(a,n) \
504    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
505    
506  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #else /* COMPILE_PCRE8 */
507    
508    #ifdef COMPILE_PCRE16
509    
510  /* When UTF-8 encoding is being used, a character is no longer just a single  #define IMM2_SIZE 1
511  byte. The macros for character handling generate simple sequences when used in  
512  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  #define PUT2(a,n,d)   \
513  never be called in byte mode. To make sure it can never even appear when UTF-8     a[n] = d
514  support is omitted, we don't even define it. */  
515    #define GET2(a,n) \
516       a[n]
517    
518    #else
519    #error Unsupported compiling mode
520    #endif /* COMPILE_PCRE16 */
521    
522  #ifndef SUPPORT_UTF8  #endif /* COMPILE_PCRE8 */
523    
524    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
525    
526    /* The maximum length of a MARK name is currently one data unit; it may be
527    changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
528    
529    #define MAX_MARK ((1 << (sizeof(pcre_uchar)*8)) - 1)
530    
531    /* When UTF encoding is being used, a character is no longer just a single
532    byte. The macros for character handling generate simple sequences when used in
533    character-mode, and more complicated ones for UTF characters. GETCHARLENTEST
534    and other macros are not used when UTF is not supported, so they are not
535    defined. To make sure they can never even appear when UTF support is omitted,
536    we don't even define them. */
537    
538    #ifndef SUPPORT_UTF
539    
540    /* #define MAX_VALUE_FOR_SINGLE_CHAR */
541    /* #define HAS_EXTRALEN(c) */
542    /* #define GET_EXTRALEN(c) */
543    /* #define NOT_FIRSTCHAR(c) */
544  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
545  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
546  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
547  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
548  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
549    /* #define GETCHARLENTEST(c, eptr, len) */
550  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
551    /* #define FORWARDCHAR(eptr) */
552    /* #define ACROSSCHAR(condition, eptr, action) */
553    
554    #else   /* SUPPORT_UTF */
555    
556    #ifdef COMPILE_PCRE8
557    
558    /* These macros were originally written in the form of loops that used data
559    from the tables whose names start with PRIV(utf8_table). They were rewritten by
560    a user so as not to use loops, because in some environments this gives a
561    significant performance advantage, and it seems never to do any harm. */
562    
563    /* Tells the biggest code point which can be encoded as a single character. */
564    
565    #define MAX_VALUE_FOR_SINGLE_CHAR 127
566    
567    /* Tests whether the code point needs extra characters to decode. */
568    
569    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
570    
571    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
572    Otherwise it has an undefined behaviour. */
573    
574    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
575    
576    /* Returns TRUE, if the given character is not the first character
577    of a UTF sequence. */
578    
579    #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
580    
581  #else   /* SUPPORT_UTF8 */  /* Base macro to pick up the remaining bytes of a UTF-8 character, not
582    advancing the pointer. */
583    
584    #define GETUTF8(c, eptr) \
585        { \
586        if ((c & 0x20) == 0) \
587          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
588        else if ((c & 0x10) == 0) \
589          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
590        else if ((c & 0x08) == 0) \
591          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
592          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
593        else if ((c & 0x04) == 0) \
594          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
595              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
596              (eptr[4] & 0x3f); \
597        else \
598          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
599              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
600              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
601        }
602    
603  /* 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
604  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
605    
606  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
607    c = *eptr; \    c = *eptr; \
608    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; \  
       } \  
     }  
609    
610  /* 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
611  pointer. */  pointer. */
612    
613  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
614    c = *eptr; \    c = *eptr; \
615    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8(c, eptr);
616    
617    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
618    the pointer. */
619    
620    #define GETUTF8INC(c, eptr) \
621      { \      { \
622      int gcii; \      if ((c & 0x20) == 0) \
623      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
624      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
625      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
626      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
627          eptr += 2; \
628          } \
629        else if ((c & 0x08) == 0) \
630          { \
631          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
632              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
633          eptr += 3; \
634          } \
635        else if ((c & 0x04) == 0) \
636          { \
637          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
638              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
639              (eptr[3] & 0x3f); \
640          eptr += 4; \
641          } \
642        else \
643        { \        { \
644        gcss -= 6; \        c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
645        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
646              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
647          eptr += 5; \
648        } \        } \
649      }      }
650    
# Line 445  know we are in UTF-8 mode. */ Line 653  know we are in UTF-8 mode. */
653    
654  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
655    c = *eptr++; \    c = *eptr++; \
656    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; \  
       } \  
     }  
657    
658  /* 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.
659    This is called when we don't know if we are in UTF-8 mode. */
660    
661  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
662    c = *eptr++; \    c = *eptr++; \
663    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
664    
665    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
666    advancing the pointer, incrementing the length. */
667    
668    #define GETUTF8LEN(c, eptr, len) \
669      { \      { \
670      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
671      int gcss = 6*gcaa; \        { \
672      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
673      while (gcaa-- > 0) \        len++; \
674          } \
675        else if ((c & 0x10)  == 0) \
676          { \
677          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
678          len += 2; \
679          } \
680        else if ((c & 0x08)  == 0) \
681          {\
682          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
683              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
684          len += 3; \
685          } \
686        else if ((c & 0x04)  == 0) \
687        { \        { \
688        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
689        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
690              (eptr[4] & 0x3f); \
691          len += 4; \
692          } \
693        else \
694          {\
695          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
696              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
697              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
698          len += 5; \
699        } \        } \
700      }      }
701    
# Line 478  if there are extra bytes. This is called Line 704  if there are extra bytes. This is called
704    
705  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
706    c = *eptr; \    c = *eptr; \
707    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
708    
709  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
710  pointer, incrementing length if there are extra bytes. This is called when we  pointer, incrementing length if there are extra bytes. This is called when we
711  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
712    
713  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
714    c = *eptr; \    c = *eptr; \
715    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
716    
717  /* 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
718  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 518  because almost all calls are already wit Line 720  because almost all calls are already wit
720    
721  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
722    
723  #endif  /* Same as above, just in the other direction. */
724    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
725    
726    /* Same as above, but it allows a fully customizable form. */
727    #define ACROSSCHAR(condition, eptr, action) \
728      while((condition) && ((eptr) & 0xc0) == 0x80) action
729    
730    #else /* COMPILE_PCRE8 */
731    
732    #ifdef COMPILE_PCRE16
733    
734    /* Tells the biggest code point which can be encoded as a single character. */
735    
736    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
737    
738    /* Tests whether the code point needs extra characters to decode. */
739    
740    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
741    
742    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
743    Otherwise it has an undefined behaviour. */
744    
745    #define GET_EXTRALEN(c) 1
746    
747    /* Returns TRUE, if the given character is not the first character
748    of a UTF sequence. */
749    
750    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
751    
752    /* Base macro to pick up the low surrogate of a UTF-16 character, not
753    advancing the pointer. */
754    
755    #define GETUTF16(c, eptr) \
756       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
757    
758    /* Get the next UTF-16 character, not advancing the pointer. This is called when
759    we know we are in UTF-16 mode. */
760    
761    #define GETCHAR(c, eptr) \
762      c = *eptr; \
763      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
764    
765    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
766    pointer. */
767    
768    #define GETCHARTEST(c, eptr) \
769      c = *eptr; \
770      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
771    
772    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
773    the pointer. */
774    
775    #define GETUTF16INC(c, eptr) \
776       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
777    
778    /* Get the next UTF-16 character, advancing the pointer. This is called when we
779    know we are in UTF-16 mode. */
780    
781    #define GETCHARINC(c, eptr) \
782      c = *eptr++; \
783      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
784    
785    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
786    This is called when we don't know if we are in UTF-16 mode. */
787    
788    #define GETCHARINCTEST(c, eptr) \
789      c = *eptr++; \
790      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
791    
792    /* Base macro to pick up the low surrogate of a UTF-16 character, not
793    advancing the pointer, incrementing the length. */
794    
795    #define GETUTF16LEN(c, eptr, len) \
796       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
797    
798    /* Get the next UTF-16 character, not advancing the pointer, incrementing
799    length if there is a low surrogate. This is called when we know we are in
800    UTF-16 mode. */
801    
802    #define GETCHARLEN(c, eptr, len) \
803      c = *eptr; \
804      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
805    
806    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
807    pointer, incrementing length if there is a low surrogate. This is called when
808    we do not know if we are in UTF-16 mode. */
809    
810    #define GETCHARLENTEST(c, eptr, len) \
811      c = *eptr; \
812      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
813    
814    /* If the pointer is not at the start of a character, move it back until
815    it is. This is called only in UTF-16 mode - we don't put a test within the
816    macro because almost all calls are already within a block of UTF-16 only
817    code. */
818    
819    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
820    
821    /* Same as above, just in the other direction. */
822    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
823    
824    /* Same as above, but it allows a fully customizable form. */
825    #define ACROSSCHAR(condition, eptr, action) \
826      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
827    
828    #endif
829    
830    #endif /* COMPILE_PCRE8 */
831    
832    #endif  /* SUPPORT_UTF */
833    
834    
835    /* Tests for Unicode horizontal and vertical whitespace characters must check a
836    number of different values. Using a switch statement for this generates the
837    fastest code (no loop, no memory access), and there are several places where
838    this happens. In order to ensure that all the case lists remain in step, we use
839    macros so that there is only one place where the lists are defined.
840    
841    NOTE: These values are also used explicitly in pcre_compile.c when processing
842    \h, \H, \v and \V in a character class, so any changes here should be
843    duplicated there as well. They also appear in pcre_jit_compile.c. */
844    
845    /* ------ ASCII/Unicode environments ------ */
846    
847    #ifndef EBCDIC
848    #define HSPACE_MULTIBYTE_CASES \
849          case 0x1680:    /* OGHAM SPACE MARK */ \
850          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */ \
851          case 0x2000:    /* EN QUAD */ \
852          case 0x2001:    /* EM QUAD */ \
853          case 0x2002:    /* EN SPACE */ \
854          case 0x2003:    /* EM SPACE */ \
855          case 0x2004:    /* THREE-PER-EM SPACE */ \
856          case 0x2005:    /* FOUR-PER-EM SPACE */ \
857          case 0x2006:    /* SIX-PER-EM SPACE */ \
858          case 0x2007:    /* FIGURE SPACE */ \
859          case 0x2008:    /* PUNCTUATION SPACE */ \
860          case 0x2009:    /* THIN SPACE */ \
861          case 0x200A:    /* HAIR SPACE */ \
862          case 0x202f:    /* NARROW NO-BREAK SPACE */ \
863          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */ \
864          case 0x3000     /* IDEOGRAPHIC SPACE */
865    
866    #define HSPACE_BYTE_CASES \
867          case CHAR_HT: \
868          case CHAR_SPACE: \
869          case 0xa0       /* NBSP */
870    
871    #define VSPACE_MULTIBYTE_CASES \
872          case 0x2028:    /* LINE SEPARATOR */ \
873          case 0x2029     /* PARAGRAPH SEPARATOR */
874    
875    #define VSPACE_BYTE_CASES \
876          case CHAR_LF: \
877          case CHAR_VT: \
878          case CHAR_FF: \
879          case CHAR_CR: \
880          case CHAR_NEL
881    
882    #define HSPACE_CASES \
883            HSPACE_BYTE_CASES: \
884            HSPACE_MULTIBYTE_CASES
885    
886    #define VSPACE_CASES \
887            VSPACE_BYTE_CASES: \
888            VSPACE_MULTIBYTE_CASES
889    
890    /* ------ EBCDIC environments ------ */
891    
892    #else
893    #define HSPACE_BYTE_CASES \
894          case CHAR_HT: \
895          case CHAR_SPACE
896    
897    #define VSPACE_BYTE_CASES \
898          case CHAR_LF: \
899          case CHAR_VT: \
900          case CHAR_FF: \
901          case CHAR_CR: \
902          case CHAR_NEL
903    
904    #define HSPACE_CASES HSPACE_BYTE_CASES
905    #define VSPACE_CASES VSPACE_BYTE_CASES
906    #endif  /* EBCDIC */
907    
908    /* ------ End of whitespace case macros ------ */
909    
910    
911  /* 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 529  Standard C system should have one. */ Line 916  Standard C system should have one. */
916  #endif  #endif
917    
918    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
919  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
920  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
921  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
922    the restrictions on partial matching have been lifted. It remains for backwards
923    compatibility. */
924    
925    #ifdef COMPILE_PCRE8
926    #define PCRE_MODE          0x0001  /* compiled in 8 bit mode */
927    #endif
928    #ifdef COMPILE_PCRE16
929    #define PCRE_MODE          0x0002  /* compiled in 16 bit mode */
930    #endif
931    #define PCRE_FIRSTSET      0x0010  /* first_char is set */
932    #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
933    #define PCRE_REQCHSET      0x0040  /* req_byte is set */
934    #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
935    #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
936    #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
937    #define PCRE_JCHANGED      0x0400  /* j option used in regex */
938    #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
939    #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
940    
941  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  /* Flags for the "extra" block produced by pcre_study(). */
 #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  
 #define PCRE_REQCHSET      0x0004  /* req_byte is set */  
 #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  
 #define PCRE_JCHANGED      0x0010  /* j option used in regex */  
 #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  
942    
943  /* Options for the "extra" block produced by pcre_study(). */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
944    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
 #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  
945    
946  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
947  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 559  time, run time, or study time, respectiv Line 954  time, run time, or study time, respectiv
954     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
955     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| \
956     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
957     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
958    
959  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
960    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
961     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
962     PCRE_NO_START_OPTIMIZE)     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
963    
964  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
965    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
966     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
967     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
968       PCRE_NO_START_OPTIMIZE)
969    
970  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
971       (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
972        PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
973    
974  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. */
 to detect whether a pattern was compiled on a host of different endianness. */  
975    
976  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
977    
978    /* This variable is used to detect a loaded regular expression
979    in different endianness. */
980    
981    #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
982    
983  /* Negative values for the firstchar and reqchar variables */  /* Negative values for the firstchar and reqchar variables */
984    
985  #define REQ_UNSET (-2)  #define REQ_UNSET (-2)
# Line 588  req_byte match. */ Line 990  req_byte match. */
990    
991  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
992    
 /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  
 variable-length repeat, or a anything other than literal characters. */  
   
 #define REQ_CASELESS 0x0100    /* indicates caselessness */  
 #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  
   
993  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
994  environments where these macros are defined elsewhere. Unfortunately, there  environments where these macros are defined elsewhere. Unfortunately, there
995  is no way to do the same for the typedef. */  is no way to do the same for the typedef. */
# Line 622  for) in a minority area (EBCDIC platform Line 1018  for) in a minority area (EBCDIC platform
1018  application that did need both could compile two versions of the library, using  application that did need both could compile two versions of the library, using
1019  macros to give the functions distinct names. */  macros to give the functions distinct names. */
1020    
1021  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF
1022    
1023  /* UTF-8 support is not enabled; use the platform-dependent character literals  /* UTF-8 support is not enabled; use the platform-dependent character literals
1024  so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */  so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF
1025    mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
1026    characters, a common practice has been to use its NL (0x15) character as the
1027    line terminator in C-like processing environments. However, sometimes the LF
1028    (0x25) character is used instead, according to this Unicode document:
1029    
1030    http://unicode.org/standard/reports/tr13/tr13-5.html
1031    
1032    PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
1033    instead. Whichever is *not* chosen is defined as NEL.
1034    
1035    In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
1036    same code point. */
1037    
1038    #ifdef EBCDIC
1039    
1040    #ifndef EBCDIC_NL25
1041    #define CHAR_NL                     '\x15'
1042    #define CHAR_NEL                    '\x25'
1043    #define STR_NL                      "\x15"
1044    #define STR_NEL                     "\x25"
1045    #else
1046    #define CHAR_NL                     '\x25'
1047    #define CHAR_NEL                    '\x15'
1048    #define STR_NL                      "\x25"
1049    #define STR_NEL                     "\x15"
1050    #endif
1051    
1052    #define CHAR_LF                     CHAR_NL
1053    #define STR_LF                      STR_NL
1054    
1055    #define CHAR_ESC                    '\047'
1056    #define CHAR_DEL                    '\007'
1057    #define STR_ESC                     "\047"
1058    #define STR_DEL                     "\007"
1059    
1060    #else  /* Not EBCDIC */
1061    
1062    /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
1063    compatibility. NEL is the Unicode newline character; make sure it is
1064    a positive value. */
1065    
1066    #define CHAR_LF                     '\n'
1067    #define CHAR_NL                     CHAR_LF
1068    #define CHAR_NEL                    ((unsigned char)'\x85')
1069    #define CHAR_ESC                    '\033'
1070    #define CHAR_DEL                    '\177'
1071    
1072    #define STR_LF                      "\n"
1073    #define STR_NL                      STR_LF
1074    #define STR_NEL                     "\x85"
1075    #define STR_ESC                     "\033"
1076    #define STR_DEL                     "\177"
1077    
1078    #endif  /* EBCDIC */
1079    
1080    /* The remaining definitions work in both environments. */
1081    
1082  #define CHAR_HT                     '\t'  #define CHAR_HT                     '\t'
1083  #define CHAR_VT                     '\v'  #define CHAR_VT                     '\v'
1084  #define CHAR_FF                     '\f'  #define CHAR_FF                     '\f'
1085  #define CHAR_CR                     '\r'  #define CHAR_CR                     '\r'
 #define CHAR_NL                     '\n'  
1086  #define CHAR_BS                     '\b'  #define CHAR_BS                     '\b'
1087  #define CHAR_BEL                    '\a'  #define CHAR_BEL                    '\a'
 #ifdef EBCDIC  
 #define CHAR_ESC                    '\047'  
 #define CHAR_DEL                    '\007'  
 #else  
 #define CHAR_ESC                    '\033'  
 #define CHAR_DEL                    '\177'  
 #endif  
1088    
1089  #define CHAR_SPACE                  ' '  #define CHAR_SPACE                  ' '
1090  #define CHAR_EXCLAMATION_MARK       '!'  #define CHAR_EXCLAMATION_MARK       '!'
# Line 742  so that PCRE works on both ASCII and EBC Line 1186  so that PCRE works on both ASCII and EBC
1186  #define STR_VT                      "\v"  #define STR_VT                      "\v"
1187  #define STR_FF                      "\f"  #define STR_FF                      "\f"
1188  #define STR_CR                      "\r"  #define STR_CR                      "\r"
 #define STR_NL                      "\n"  
1189  #define STR_BS                      "\b"  #define STR_BS                      "\b"
1190  #define STR_BEL                     "\a"  #define STR_BEL                     "\a"
 #ifdef EBCDIC  
 #define STR_ESC                     "\047"  
 #define STR_DEL                     "\007"  
 #else  
 #define STR_ESC                     "\033"  
 #define STR_DEL                     "\177"  
 #endif  
1191    
1192  #define STR_SPACE                   " "  #define STR_SPACE                   " "
1193  #define STR_EXCLAMATION_MARK        "!"  #define STR_EXCLAMATION_MARK        "!"
# Line 853  so that PCRE works on both ASCII and EBC Line 1289  so that PCRE works on both ASCII and EBC
1289  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
1290  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
1291  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
1292    #define STRING_MARK0                "MARK\0"
1293  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
1294  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
1295  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 874  so that PCRE works on both ASCII and EBC Line 1311  so that PCRE works on both ASCII and EBC
1311    
1312  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
1313    
1314  #define STRING_CR_RIGHTPAR          "CR)"  #define STRING_CR_RIGHTPAR             "CR)"
1315  #define STRING_LF_RIGHTPAR          "LF)"  #define STRING_LF_RIGHTPAR             "LF)"
1316  #define STRING_CRLF_RIGHTPAR        "CRLF)"  #define STRING_CRLF_RIGHTPAR           "CRLF)"
1317  #define STRING_ANY_RIGHTPAR         "ANY)"  #define STRING_ANY_RIGHTPAR            "ANY)"
1318  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1319  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1320  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1321  #define STRING_UTF8_RIGHTPAR        "UTF8)"  #ifdef COMPILE_PCRE8
1322    #define STRING_UTF_RIGHTPAR            "UTF8)"
1323    #endif
1324    #ifdef COMPILE_PCRE16
1325    #define STRING_UTF_RIGHTPAR            "UTF16)"
1326    #endif
1327    #define STRING_UCP_RIGHTPAR            "UCP)"
1328    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1329    
1330  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF */
1331    
1332  /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This  /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1333  works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode  works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
# Line 893  only. */ Line 1337  only. */
1337  #define CHAR_VT                     '\013'  #define CHAR_VT                     '\013'
1338  #define CHAR_FF                     '\014'  #define CHAR_FF                     '\014'
1339  #define CHAR_CR                     '\015'  #define CHAR_CR                     '\015'
1340  #define CHAR_NL                     '\012'  #define CHAR_LF                     '\012'
1341    #define CHAR_NL                     CHAR_LF
1342    #define CHAR_NEL                    ((unsigned char)'\x85')
1343  #define CHAR_BS                     '\010'  #define CHAR_BS                     '\010'
1344  #define CHAR_BEL                    '\007'  #define CHAR_BEL                    '\007'
1345  #define CHAR_ESC                    '\033'  #define CHAR_ESC                    '\033'
# Line 1105  only. */ Line 1551  only. */
1551  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1552  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1553  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1554    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1555  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1556  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1557  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1126  only. */ Line 1573  only. */
1573    
1574  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1575    
1576  #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1577  #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1578  #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1579  #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1580  #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1581  #define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1582  #define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1583  #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #ifdef COMPILE_PCRE8
1584    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1585    #endif
1586    #ifdef COMPILE_PCRE16
1587    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1588    #endif
1589    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1590    #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
1591    
1592  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
1593    
1594  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1595    
# Line 1148  only. */ Line 1602  only. */
1602  #endif  #endif
1603    
1604  #ifndef ESC_n  #ifndef ESC_n
1605  #define ESC_n CHAR_NL  #define ESC_n CHAR_LF
1606  #endif  #endif
1607    
1608  #ifndef ESC_r  #ifndef ESC_r
# Line 1166  only. */ Line 1620  only. */
1620    
1621  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1622  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1623  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1624  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1625  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1626    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1627    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1628    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1629    #define PT_WORD       8    /* Word - L plus N plus underscore */
1630    
1631  /* 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
1632  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1633    
1634  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1635  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 1185  contain UTF-8 characters with values gre Line 1643  contain UTF-8 characters with values gre
1643  /* 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
1644  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
1645  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
1646  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
1647  corresponds to "." rather than an escape sequence, and another for OP_ALLANY  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1648  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1649    non-DOTALL mode, "." behaves like \N.
1650    
1651    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1652    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1653    They must be contiguous, and remain in order so that the replacements can be
1654    looked up from a table.
1655    
1656  The final escape must be ESC_REF as subsequent values are used for  The final escape must be ESC_REF as subsequent values are used for
1657  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
# Line 1197  put in between that don't consume a char Line 1661  put in between that don't consume a char
1661  */  */
1662    
1663  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,
1664         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1665         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1666           ESC_E, ESC_Q, ESC_g, ESC_k,
1667           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1668         ESC_REF };         ESC_REF };
1669    
   
1670  /* 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
1671  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.
1672    
1673  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1674  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
1675  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1676    
1677  enum {  enum {
1678    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1225  enum { Line 1690  enum {
1690    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1691    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1692    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1693    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1694    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1695      OP_ALLANY,         /* 13 Match any character */
1696    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1697    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1698    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1239  enum { Line 1705  enum {
1705    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1706    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1707    
1708    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1709    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1710    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1711    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1712    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1713    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1714      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1715    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1716    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1717    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1718    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1719    OP_QUERY,          /* 35 */  
1720    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1721    
1722    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1723    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1724    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1725      OP_MINPLUS,        /* 36 */
1726    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1727    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1728    OP_POSQUERY,       /* 42 Posesssified query */  
1729    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1730      OP_MINUPTO,        /* 40 */
1731    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1732    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1733    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1734    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1735    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1736    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1737    
1738    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1739    OP_NOTMINUPTO,     /* 51 */  
1740    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1741      OP_MINSTARI,       /* 47 */
1742    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1743    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1744    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1745    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1746    
1747    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1748    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1749    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1750    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1751    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1752    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1753      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1754    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1755    OP_TYPEMINUPTO,    /* 64 */  
1756    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1757      /**** Negated single character, caseful; must precede the caseless ones ****/
1758    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1759    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1760    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1761    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1762      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1763    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1764    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1765    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1766    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1767    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1768    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1769    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1770    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1771      OP_NOTPOSPLUS,     /* 69 */
1772    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1773    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1774                             class - the difference is relevant only when a UTF-8  
1775                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1776    
1777    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1778                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1779      OP_NOTPLUSI,       /* 74 */
1780    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1781    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1782    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1783    
1784    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1785    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1786    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1787    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1788      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1789    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1790      OP_NOTPOSQUERYI,   /* 83 */
1791    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1792    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1793    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1794    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1795    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1796      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1797    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1798    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1799      OP_TYPEQUERY,      /* 89 */
1800    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1801    OP_BRA,            /* 94 Start of non-capturing bracket */  
1802    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1803    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1804      OP_TYPEEXACT,      /* 93 Exactly n matches */
1805    
1806      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1807      OP_TYPEPOSPLUS,    /* 95 */
1808      OP_TYPEPOSQUERY,   /* 96 */
1809      OP_TYPEPOSUPTO,    /* 97 */
1810    
1811      /* These are used for character classes and back references; only the
1812      first six are the same as the sets above. */
1813    
1814      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1815      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1816      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1817      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1818      OP_CRQUERY,        /* 102 */
1819      OP_CRMINQUERY,     /* 103 */
1820    
1821      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1822      OP_CRMINRANGE,     /* 105 */
1823    
1824      /* End of quantifier opcodes */
1825    
1826      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1827      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1828                                  class - the difference is relevant only when a
1829                                  character > 255 is encountered. */
1830      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1831                                  class. This does both positive and negative. */
1832      OP_REF,            /* 109 Match a back reference, casefully */
1833      OP_REFI,           /* 110 Match a back reference, caselessly */
1834      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1835      OP_CALLOUT,        /* 112 Call out to external function if provided */
1836    
1837      OP_ALT,            /* 113 Start of alternation */
1838      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1839      OP_KETRMAX,        /* 115 These two must remain together and in this */
1840      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1841      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1842    
1843      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1844      asserts must remain in order. */
1845    
1846      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1847      OP_ASSERT,         /* 119 Positive lookahead */
1848      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1849      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1850      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1851    
1852      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1853      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1854      subpattern that isn't an assertion. The POS versions must immediately follow
1855      the non-POS versions in each case. */
1856    
1857      OP_ONCE,           /* 123 Atomic group, contains captures */
1858      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1859      OP_BRA,            /* 125 Start of non-capturing bracket */
1860      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1861      OP_CBRA,           /* 127 Start of capturing bracket */
1862      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1863      OP_COND,           /* 129 Conditional group */
1864    
1865    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1866    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1867    
1868    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1869    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1870    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1871      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1872    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_SCOND,          /* 134 Conditional group, check empty */
1873    OP_RREF,           /* 101 Used to hold a recursion number as condition */  
1874    OP_DEF,            /* 102 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1875    
1876    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_CREF,           /* 135 Used to hold a capture number as condition */
1877    OP_BRAMINZERO,     /* 104 order. */    OP_NCREF,          /* 136 Same, but generated by a name reference*/
1878      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1879      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1880      OP_DEF,            /* 139 The DEFINE condition */
1881    
1882      OP_BRAZERO,        /* 140 These two must remain together and in this */
1883      OP_BRAMINZERO,     /* 141 order. */
1884      OP_BRAPOSZERO,     /* 142 */
1885    
1886    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1887    
1888    OP_PRUNE,          /* 105 */    OP_MARK,           /* 143 always has an argument */
1889    OP_SKIP,           /* 106 */    OP_PRUNE,          /* 144 */
1890    OP_THEN,           /* 107 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1891    OP_COMMIT,         /* 108 */    OP_SKIP,           /* 146 */
1892      OP_SKIP_ARG,       /* 147 same, but with argument */
1893      OP_THEN,           /* 148 */
1894      OP_THEN_ARG,       /* 149 same, but with argument */
1895      OP_COMMIT,         /* 150 */
1896    
1897    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1898    
1899    OP_FAIL,           /* 109 */    OP_FAIL,           /* 151 */
1900    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 152 */
1901      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1902      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1903    
1904    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1905    
1906    OP_SKIPZERO        /* 111 */    OP_SKIPZERO,       /* 155 */
1907    
1908      /* This is not an opcode, but is used to check that tables indexed by opcode
1909      are the correct length, in order to catch updating errors - there have been
1910      some in the past. */
1911    
1912      OP_TABLE_LENGTH
1913  };  };
1914    
1915    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1916    definitions that follow must also be updated to match. There are also tables
1917    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1918    
1919    
1920  /* 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
1921  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
1922    only in pcre_printint.c, which fills out the full names in many cases (and in
1923    some cases doesn't actually use these names at all). */
1924    
1925  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1926    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1927    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1928    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1929    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1930    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1931    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1932      "{", "{", "{",                                                  \
1933    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1934    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1935      "{", "{", "{",                                                  \
1936      "*+","++", "?+", "{",                                           \
1937      "*", "*?", "+", "+?", "?", "??",                                \
1938      "{", "{", "{",                                                  \
1939      "*+","++", "?+", "{",                                           \
1940      "*", "*?", "+", "+?", "?", "??",                                \
1941      "{", "{", "{",                                                  \
1942    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1943    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1944    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1945    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1946    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1947    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1948    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1949    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1950    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once", "Once_NC",                                              \
1951    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Bra", "BraPos", "CBra", "CBraPos",                             \
1952    "Skip zero"    "Cond",                                                         \
1953      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1954      "SCond",                                                        \
1955      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1956      "Brazero", "Braminzero", "Braposzero",                          \
1957      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1958      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1959      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1960      "Close", "Skip zero"
1961    
1962    
1963  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 1407  in UTF-8 mode. The code that uses this t Line 1974  in UTF-8 mode. The code that uses this t
1974    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1975    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1976    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1977    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1978    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1979    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1980      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1981    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1982    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1983    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1984    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1985    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1986    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1987    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1988      2+IMM2_SIZE,                   /* exact                                  */ \
1989      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1990      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1991      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1992      2+IMM2_SIZE,                   /* exact I                                */ \
1993      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1994    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1995    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1996    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1997    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1998      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1999      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
2000      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
2001      2+IMM2_SIZE,                   /* NOT exact I                            */ \
2002      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
2003    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
2004    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
2005    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
2006    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
2007      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
2008    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2009    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2010    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2011   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2012   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2013    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2014    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2015      1+IMM2_SIZE,                   /* REFI                                   */ \
2016    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2017    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2018    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
2019    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
2020    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
2021    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
2022      1+LINK_SIZE,                   /* KetRpos                                */ \
2023      1+LINK_SIZE,                   /* Reverse                                */ \
2024    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
2025    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
2026    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
2027    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
2028    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
2029      1+LINK_SIZE,                   /* ONCE_NC                                */ \
2030    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
2031    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
2032      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
2033      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
2034    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
2035    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
2036    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
2037      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2038      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2039    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2040    3,                             /* CREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
2041    3,                             /* RREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
2042    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2043    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2044    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
2045    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
2046      1, 3,                          /* THEN, THEN_ARG                         */ \
2047      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2048      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2049    
2050  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
2051    condition. */
2052    
2053  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
2054    
2055  /* 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
2056  tracked. */  easily be tracked. When a new number is added, the table called eint in
2057    pcreposix.c must be updated. */
2058    
2059  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
2060         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1471  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2062  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2062         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
2063         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2064         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2065         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2066           ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERRCOUNT };
2067    
2068    /* JIT compiling modes. The function list is indexed by them. */
2069    enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2070           JIT_NUMBER_OF_COMPILE_MODES };
2071    
2072  /* 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
2073  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 1487  Because people can now save and re-use c Line 2083  Because people can now save and re-use c
2083  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
2084  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
2085  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
2086  NOTE NOTE NOTE:  NOTE NOTE NOTE
2087  */  */
2088    
2089  typedef struct real_pcre {  #ifdef COMPILE_PCRE8
2090    #define REAL_PCRE real_pcre
2091    #else
2092    #define REAL_PCRE real_pcre16
2093    #endif
2094    
2095    typedef struct REAL_PCRE {
2096    pcre_uint32 magic_number;    pcre_uint32 magic_number;
2097    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2098    pcre_uint32 options;            /* Public options */    pcre_uint32 options;            /* Public options */
2099    pcre_uint16 flags;              /* Private flags */    pcre_uint16 flags;              /* Private flags */
2100    pcre_uint16 dummy1;             /* For future use */    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2101    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;        /* Highest numbered group */
2102    pcre_uint16 top_backref;    pcre_uint16 top_backref;        /* Highest numbered back reference */
2103    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
2104    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
2105    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2106    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
2107    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
2108    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
2109      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2110    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *nullpad;      /* NULL padding */
2111    const unsigned char *nullpad;   /* NULL padding */  } REAL_PCRE;
 } real_pcre;  
2112    
2113  /* 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
2114  remark (see NOTE above) about extending this structure applies. */  remark (see NOTE above) about extending this structure applies. */
2115    
2116  typedef struct pcre_study_data {  typedef struct pcre_study_data {
2117    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2118    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
2119    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
2120      pcre_uint32 minlength;          /* Minimum subject length */
2121  } pcre_study_data;  } pcre_study_data;
2122    
2123    /* Structure for building a chain of open capturing subpatterns during
2124    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2125    encountered. This is also used to identify subpatterns that contain recursive
2126    back references to themselves, so that they can be made atomic. */
2127    
2128    typedef struct open_capitem {
2129      struct open_capitem *next;    /* Chain link */
2130      pcre_uint16 number;           /* Capture number */
2131      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2132    } open_capitem;
2133    
2134  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2135  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2136    
2137  typedef struct compile_data {  typedef struct compile_data {
2138    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2139    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2140    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2141    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2142    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
2143    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2144    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2145    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2146    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
2147    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
2148    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
2149    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
2150    int  bracount;                /* Count of capturing parens as we compile */    int  name_entry_size;             /* Size of each entry */
2151    int  final_bracount;          /* Saved value after first pass */    int  workspace_size;              /* Size of workspace */
2152    int  top_backref;             /* Maximum back reference */    int  bracount;                    /* Count of capturing parens as we compile */
2153    unsigned int backref_map;     /* Bitmap of low back refs */    int  final_bracount;              /* Saved value after first pass */
2154    int  external_options;        /* External (initial) options */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2155    int  external_flags;          /* External flag bits to be set */    int  top_backref;                 /* Maximum back reference */
2156    int  req_varyopt;             /* "After variable item" flag for reqbyte */    unsigned int backref_map;         /* Bitmap of low back refs */
2157    BOOL had_accept;              /* (*ACCEPT) encountered */    int  assert_depth;                /* Depth of nested assertions */
2158    int  nltype;                  /* Newline type */    int  external_options;            /* External (initial) options */
2159    int  nllen;                   /* Newline string length */    int  external_flags;              /* External flag bits to be set */
2160    uschar nl[4];                 /* Newline string when fixed length */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2161      BOOL had_accept;                  /* (*ACCEPT) encountered */
2162      BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2163      BOOL check_lookbehind;            /* Lookbehinds need later checking */
2164      int  nltype;                      /* Newline type */
2165      int  nllen;                       /* Newline string length */
2166      pcre_uchar nl[4];                 /* Newline string when fixed length */
2167  } compile_data;  } compile_data;
2168    
2169  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
2170  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
2171    
2172  typedef struct branch_chain {  typedef struct branch_chain {
2173    struct branch_chain *outer;    struct branch_chain *outer;
2174    uschar *current;    pcre_uchar *current_branch;
2175  } branch_chain;  } branch_chain;
2176    
2177  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2178  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2179    
2180  typedef struct recursion_info {  typedef struct recursion_info {
2181    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2182    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2183    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2184    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
2185    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 */  
2186  } recursion_info;  } recursion_info;
2187    
2188    /* A similar structure for pcre_dfa_exec(). */
2189    
2190    typedef struct dfa_recursion_info {
2191      struct dfa_recursion_info *prevrec;
2192      int group_num;
2193      PCRE_PUCHAR subject_position;
2194    } dfa_recursion_info;
2195    
2196  /* 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
2197  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
2198  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
2199    pcre_exec(). */
2200    
2201  typedef struct eptrblock {  typedef struct eptrblock {
2202    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2203    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2204  } eptrblock;  } eptrblock;
2205    
2206    
# Line 1584  typedef struct match_data { Line 2211  typedef struct match_data {
2211    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2212    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2213    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2214    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2215    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2216    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2217    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2218    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2219    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
2220    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
2221    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
2222    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
2223    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2224    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2225    BOOL   utf8;                  /* UTF8 flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2226    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   offset_overflow;         /* Set if too many extractions */
2227    BOOL   endonly;               /* Dollar not before final \n */    BOOL   notbol;                  /* NOTBOL flag */
2228    BOOL   notempty;              /* Empty string match not wanted */    BOOL   noteol;                  /* NOTEOL flag */
2229    BOOL   partial;               /* PARTIAL flag */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2230    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2231    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2232    const uschar *start_code;     /* For use when recursing */    BOOL   endonly;                 /* Dollar not before final \n */
2233    USPTR  start_subject;         /* Start of the subject string */    BOOL   notempty;                /* Empty string match not wanted */
2234    USPTR  end_subject;           /* End of the subject string */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2235    USPTR  start_match_ptr;       /* Start of matched string */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2236    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2237    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2238    int    capture_last;          /* Most recent capture number */    BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */
2239    int    start_offset;          /* The start offset value */    const  pcre_uchar *start_code;  /* For use when recursing */
2240    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2241    int    eptrn;                 /* Next free eptrblock */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2242    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2243    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2244      PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2245      int    partial;                 /* PARTIAL options */
2246      int    end_offset_top;          /* Highwater mark at end of match */
2247      int    capture_last;            /* Most recent capture number */
2248      int    start_offset;            /* The start offset value */
2249      int    match_function_type;     /* Set for certain special calls of MATCH() */
2250      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2251      int    eptrn;                   /* Next free eptrblock */
2252      recursion_info *recursive;      /* Linked list of recursion data */
2253      void  *callout_data;            /* To pass back to callouts */
2254      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2255      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2256      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2257    #ifdef NO_RECURSE
2258      void  *match_frames_base;       /* For remembering malloc'd frames */
2259    #endif
2260  } match_data;  } match_data;
2261    
2262  /* 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
2263  functions. */  functions. */
2264    
2265  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2266    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2267    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2268    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2269    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2270    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
2271    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
2272    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
2273    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
2274    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
2275    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
2276      pcre_uchar nl[4];                 /* Newline string when fixed */
2277      void *callout_data;               /* To pass back to callouts */
2278      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2279  } dfa_match_data;  } dfa_match_data;
2280    
2281  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1665  total length. */ Line 2311  total length. */
2311  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2312  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2313    
2314    /* Internal function and data prefixes. */
2315    
2316    #ifdef COMPILE_PCRE8
2317    #ifndef PUBL
2318    #define PUBL(name) pcre_##name
2319    #endif
2320    #ifndef PRIV
2321    #define PRIV(name) _pcre_##name
2322    #endif
2323    #else /* COMPILE_PCRE8 */
2324    #ifdef COMPILE_PCRE16
2325    #ifndef PUBL
2326    #define PUBL(name) pcre16_##name
2327    #endif
2328    #ifndef PRIV
2329    #define PRIV(name) _pcre16_##name
2330    #endif
2331    #else
2332    #error Unsupported compiling mode
2333    #endif /* COMPILE_PCRE16 */
2334    #endif /* COMPILE_PCRE8 */
2335    
2336  /* 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
2337  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
2338  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 1682  of the exported public functions. They h Line 2350  of the exported public functions. They h
2350  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
2351  pcre_tables.c module. */  pcre_tables.c module. */
2352    
2353  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2354  extern const int    _pcre_utf8_table2[];  
2355  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1)[];
2356  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table1_size);
2357    extern const int            PRIV(utf8_table2)[];
2358    extern const int            PRIV(utf8_table3)[];
2359    extern const pcre_uint8     PRIV(utf8_table4)[];
2360    
2361  extern const int    _pcre_utf8_table1_size;  #endif /* COMPILE_PCRE8 */
2362    
2363  extern const char   _pcre_utt_names[];  extern const char           PRIV(utt_names)[];
2364  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table PRIV(utt)[];
2365  extern const int _pcre_utt_size;  extern const int            PRIV(utt_size);
2366    
2367  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2368    
2369  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2370    
2371    
2372  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2373  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
2374  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2375    
2376  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  /* String comparison functions. */
2377                        int *, BOOL);  #ifdef COMPILE_PCRE8
 extern int          _pcre_ord2utf8(int, uschar *);  
 extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  
                       const pcre_study_data *, pcre_study_data *);  
 extern int          _pcre_valid_utf8(const uschar *, int);  
 extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  
                       int *, BOOL);  
 extern BOOL         _pcre_xclass(int, const uschar *);  
2378    
2379    #define STRCMP_UC_UC(str1, str2) \
2380      strcmp((char *)(str1), (char *)(str2))
2381    #define STRCMP_UC_C8(str1, str2) \
2382      strcmp((char *)(str1), (str2))
2383    #define STRNCMP_UC_UC(str1, str2, num) \
2384      strncmp((char *)(str1), (char *)(str2), (num))
2385    #define STRNCMP_UC_C8(str1, str2, num) \
2386      strncmp((char *)(str1), (str2), (num))
2387    #define STRLEN_UC(str) strlen((const char *)str)
2388    
2389    #else
2390    
2391    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2392                               const pcre_uchar *);
2393    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2394                               const char *);
2395    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2396                               const pcre_uchar *, unsigned int num);
2397    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2398                               const char *, unsigned int num);
2399    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2400    
2401    #define STRCMP_UC_UC(str1, str2) \
2402      PRIV(strcmp_uc_uc)((str1), (str2))
2403    #define STRCMP_UC_C8(str1, str2) \
2404      PRIV(strcmp_uc_c8)((str1), (str2))
2405    #define STRNCMP_UC_UC(str1, str2, num) \
2406      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2407    #define STRNCMP_UC_C8(str1, str2, num) \
2408      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2409    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2410    
2411    #endif /* COMPILE_PCRE8 */
2412    
2413    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2414    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2415                               int *, BOOL);
2416    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2417    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2418    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2419                               int *, BOOL);
2420    extern BOOL              PRIV(xclass)(int, const pcre_uchar *, BOOL);
2421    
2422    #ifdef SUPPORT_JIT
2423    extern void              PRIV(jit_compile)(const REAL_PCRE *,
2424                               PUBL(extra) *, int);
2425    extern int               PRIV(jit_exec)(const REAL_PCRE *, const PUBL(extra) *,
2426                               const pcre_uchar *, int, int, int, int *, int);
2427    extern void              PRIV(jit_free)(void *);
2428    extern int               PRIV(jit_get_size)(void *);
2429    extern const char*       PRIV(jit_get_target)(void);
2430    #endif
2431    
2432  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
2433    
2434  typedef struct {  typedef struct {
2435    uschar script;    pcre_uint8 script;     /* ucp_Arabic, etc. */
2436    uschar chartype;    pcre_uint8 chartype;   /* ucp_Cc, etc. (general categories) */
2437    pcre_int32 other_case;    pcre_uint8 gbprop;     /* ucp_gbControl, etc. (grapheme break property) */
2438      pcre_uint8 caseset;    /* offset to multichar other cases or zero */
2439      pcre_int32 other_case; /* offset to other case, or zero if none */
2440  } ucd_record;  } ucd_record;
2441    
2442  extern const ucd_record  _pcre_ucd_records[];  extern const pcre_uint32 PRIV(ucd_caseless_sets)[];
2443  extern const uschar      _pcre_ucd_stage1[];  extern const ucd_record  PRIV(ucd_records)[];
2444  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint8  PRIV(ucd_stage1)[];
2445  extern const int         _pcre_ucp_gentype[];  extern const pcre_uint16 PRIV(ucd_stage2)[];
2446    extern const int         PRIV(ucp_gentype)[];
2447    extern const pcre_uint32 PRIV(ucp_gbtable)[];
2448    #ifdef SUPPORT_JIT
2449    extern const int         PRIV(ucp_typerange)[];
2450    #endif
2451    
2452    #ifdef SUPPORT_UCP
2453  /* UCD access macros */  /* UCD access macros */
2454    
2455  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
2456  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (PRIV(ucd_records) + \
2457          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2458          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2459    
2460  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)    GET_UCD(ch)->chartype
2461  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)      GET_UCD(ch)->script
2462  #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]  #define UCD_CATEGORY(ch)    PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2463  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)  #define UCD_GRAPHBREAK(ch)  GET_UCD(ch)->gbprop
2464    #define UCD_CASESET(ch)     GET_UCD(ch)->caseset
2465    #define UCD_OTHERCASE(ch)   (ch + GET_UCD(ch)->other_case)
2466    
2467    #endif /* SUPPORT_UCP */
2468    
2469  #endif  #endif
2470    

Legend:
Removed from v.412  
changed lines
  Added in v.1044

  ViewVC Help
Powered by ViewVC 1.1.5