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

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

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

code/trunk/pcre_internal.h revision 483 by ph10, Mon Jan 4 16:19:17 2010 UTC code/branches/pcre16/pcre_internal.h revision 782 by zherczeg, Sat Dec 3 23:58:37 2011 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2010 University of Cambridge             Copyright (c) 1997-2011 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 40  POSSIBILITY OF SUCH DAMAGE. Line 40  POSSIBILITY OF SUCH DAMAGE.
40    
41  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_" or "_pcre16_" depending on
44    the PRIV macro. */
45    
46  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
47  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
# Line 51  functions whose names all begin with "_p Line 52  functions whose names all begin with "_p
52  #define PCRE_DEBUG  #define PCRE_DEBUG
53  #endif  #endif
54    
55  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"  /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56    #ifndef COMPILE_PCRE16
57    #define COMPILE_PCRE8
58    #endif
59    
60    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
61  script prevents both being selected, but not everybody uses "configure". */  script prevents both being selected, but not everybody uses "configure". */
62    
63  #if defined EBCDIC && defined SUPPORT_UTF8  #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
64  #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.  #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
65  #endif  #endif
66    
67  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
# Line 65  script prevents both being selected, but Line 71  script prevents both being selected, but
71  #define SUPPORT_UTF8 1  #define SUPPORT_UTF8 1
72  #endif  #endif
73    
74    /* If SUPPORT_UCP is defined, SUPPORT_UTF16 must also be defined. The
75    "configure" script ensures this, but not everybody uses "configure". */
76    
77    #if defined SUPPORT_UCP && defined COMPILE_PCRE16 && !defined SUPPORT_UTF16
78    #define SUPPORT_UTF16 1
79    #endif
80    
81    /* This macro is defined if either UTF-8 or UTF-16 support or both are
82    enabled. */
83    
84    #if defined SUPPORT_UTF8 || defined SUPPORT_UTF16
85    /* Unicode Transformation Format is enabled. */
86    #define SUPPORT_UTF 1
87    #endif
88    
89  /* 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
90  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
91  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 158  set, we ensure here that it has no effec Line 179  set, we ensure here that it has no effec
179  #define PCRE_CALL_CONVENTION  #define PCRE_CALL_CONVENTION
180  #endif  #endif
181    
182  /* 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
183  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
184  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
185  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
186  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
187    
188    typedef unsigned char pcre_uint8;
189    
190  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
191    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
192    typedef short pcre_int16;    typedef short pcre_int16;
# Line 184  preprocessor time in standard C environm Line 207  preprocessor time in standard C environm
207    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
208  #endif  #endif
209    
210  /* When checking for integer overflow in pcre_compile(), we need to handle  /* When checking for integer overflow in pcre_compile(), we need to handle
211  large integers. If a 64-bit integer type is available, we can use that.  large integers. If a 64-bit integer type is available, we can use that.
212  Otherwise we have to cast to double, which of course requires floating point  Otherwise we have to cast to double, which of course requires floating point
213  arithmetic. Handle this by defining a macro for the appropriate type. If  arithmetic. Handle this by defining a macro for the appropriate type. If
214  stdint.h is available, include it; it may define INT64_MAX. The macro int64_t  stdint.h is available, include it; it may define INT64_MAX. Systems that do not
215  may be set by "configure". */  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
216    by "configure". */
217    
218  #if HAVE_STDINT_H  #if HAVE_STDINT_H
219  #include <stdint.h>  #include <stdint.h>
220    #elif HAVE_INTTYPES_H
221    #include <inttypes.h>
222  #endif  #endif
223    
224  #if defined INT64_MAX || defined int64_t  #if defined INT64_MAX || defined int64_t
# Line 203  may be set by "configure". */ Line 229  may be set by "configure". */
229    
230  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
231  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
232  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 *,
233  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
234  to save lots of typing. I tried "uchar", but it causes problems on Digital  called pcre_uchar.
235  Unix, where it is defined in sys/types, so use "uschar" instead. */  
236    The IN_UCHARS macro multiply its argument with the byte size of the current
237    pcre_uchar type. Useful for memcpy and such operations, whose require the
238    byte size of their input/output buffers.
239    
240    The MAX_255 macro checks whether its pcre_uchar input is less than 256.
241    
242    The TABLE_GET macro is designed for accessing elements of tables whose contain
243    exactly 256 items. When the character is able to contain more than 256
244    items, some check is needed before accessing these tables.
245    */
246    
247    #ifdef COMPILE_PCRE8
248    
249  typedef unsigned char uschar;  typedef unsigned char pcre_uchar;
250    #define IN_UCHARS(x) (x)
251    #define MAX_255(c) 1
252    #define TABLE_GET(c, table, default) ((table)[c])
253    
254    #else
255    
256    #ifdef COMPILE_PCRE16
257    #if USHRT_MAX != 65535
258    /* This is a warning message. Change PCRE_SCHAR16 to a 16 bit data type in
259    pcre.h(.in) and disable (comment out) this message. */
260    #error Warning: PCRE_SCHAR16 is not a 16 bit data type.
261    #endif
262    
263    typedef pcre_uint16 pcre_uchar;
264    #define IN_UCHARS(x) ((x) << 1)
265    #define MAX_255(c) ((c) <= 255u)
266    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
267    
268    #else
269    #error Unsupported compiling mode
270    #endif /* COMPILE_PCRE16 */
271    
272    #endif /* COMPILE_PCRE8 */
273    
274  /* 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
275  characters only go up to 0x7fffffff (though Unicode doesn't go beyond  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
# Line 231  start/end of string field names are. */ Line 292  start/end of string field names are. */
292  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
293    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
294      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
295       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
296         utf8)) \         &(NLBLOCK->nllen), utf)) \
297      : \      : \
298      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
299       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 245  start/end of string field names are. */ Line 306  start/end of string field names are. */
306  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
307    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
308      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
309       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
310         &(NLBLOCK->nllen), utf8)) \         &(NLBLOCK->nllen), utf)) \
311      : \      : \
312      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
313       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 264  used for the external interface and appe Line 325  used for the external interface and appe
325  must begin with PCRE_. */  must begin with PCRE_. */
326    
327  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
328  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
329  #else  #else
330  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
331  #endif  #endif
332    
   
   
333  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
334  values. */  values. */
335    
# Line 340  The macros are controlled by the value o Line 397  The macros are controlled by the value o
397  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
398  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
399    
400    #ifdef COMPILE_PCRE8
401    
402  #if LINK_SIZE == 2  #if LINK_SIZE == 2
403    
404  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 376  is automated on Unix systems via the "co Line 435  is automated on Unix systems via the "co
435  #define GET(a,n) \  #define GET(a,n) \
436    (((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])
437    
438  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
439    #define MAX_PATTERN_SIZE (1 << 30)
440    
441    #else
442    #error LINK_SIZE must be either 2, 3, or 4
443    #endif
444    
445    #else /* COMPILE_PCRE8 */
446    
447    #ifdef COMPILE_PCRE16
448    
449    #if LINK_SIZE == 2
450    
451    #undef LINK_SIZE
452    #define LINK_SIZE 1
453    
454    #define PUT(a,n,d)   \
455      (a[n] = (d))
456    
457    #define GET(a,n) \
458      (a[n])
459    
460    #define MAX_PATTERN_SIZE (1 << 16)
461    
462    #elif LINK_SIZE == 3 || LINK_SIZE == 4
463    
464    #undef LINK_SIZE
465    #define LINK_SIZE 2
466    
467    #define PUT(a,n,d)   \
468      (a[n] = (d) >> 16), \
469      (a[(n)+1] = (d) & 65536)
470    
471    #define GET(a,n) \
472      (((a)[n] << 16) | (a)[(n)+1])
473    
474    /* Keep it positive */
475    #define MAX_PATTERN_SIZE (1 << 30)
476    
477  #else  #else
478  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
479  #endif  #endif
480    
481    #else
482    #error Unsupported compiling mode
483    #endif /* COMPILE_PCRE16 */
484    
485    #endif /* COMPILE_PCRE8 */
486    
487  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
488    
# Line 393  is automated on Unix systems via the "co Line 493  is automated on Unix systems via the "co
493  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
494  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
495    
496    #ifdef COMPILE_PCRE8
497    
498    #define IMM2_SIZE 2
499    
500  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
501    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
502    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 400  capturing parenthesis numbers in back re Line 504  capturing parenthesis numbers in back re
504  #define GET2(a,n) \  #define GET2(a,n) \
505    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
506    
507  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #else /* COMPILE_PCRE8 */
508    
509    #ifdef COMPILE_PCRE16
510    
511    #define IMM2_SIZE 1
512    
513    #define PUT2(a,n,d)   \
514       a[n] = d
515    
516    #define GET2(a,n) \
517       a[n]
518    
519    #else
520    #error Unsupported compiling mode
521    #endif /* COMPILE_PCRE16 */
522    
523    #endif /* COMPILE_PCRE8 */
524    
525    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
526    
527  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
528  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
529  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
530  never be called in byte mode. To make sure it can never even appear when UTF-8  not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should
531  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
532    UTF-8 support is omitted, we don't even define them. */
533  #ifndef SUPPORT_UTF8  
534    /* #define HAS_EXTRALEN(c) */
535    /* #define GET_EXTRALEN(c) */
536    #ifndef SUPPORT_UTF
537  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
538  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
539  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
540  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
541  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
542    /* #define GETCHARLENTEST(c, eptr, len) */
543  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
544    /* #define FORWARDCHAR(eptr) */
545    /* #define INTERNALCHAR(condition, eptr, action) */
546    
547    #else   /* SUPPORT_UTF */
548    
549    #ifdef COMPILE_PCRE8
550    
551    /* These macros were originally written in the form of loops that used data
552    from the tables whose names start with PRIV(utf8_table). They were rewritten by
553    a user so as not to use loops, because in some environments this gives a
554    significant performance advantage, and it seems never to do any harm. */
555    
556    /* Tests whether the code point needs extra characters to decode. */
557    
558    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
559    
560    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
561    Otherwise it has an undefined behaviour. */
562    
563    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
564    
565    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
566    advancing the pointer. */
567    
568  #else   /* SUPPORT_UTF8 */  #define GETUTF8(c, eptr) \
569        { \
570        if ((c & 0x20) == 0) \
571          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
572        else if ((c & 0x10) == 0) \
573          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
574        else if ((c & 0x08) == 0) \
575          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
576          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
577        else if ((c & 0x04) == 0) \
578          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
579              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
580              (eptr[4] & 0x3f); \
581        else \
582          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
583              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
584              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
585        }
586    
587  /* 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
588  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
589    
590  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
591    c = *eptr; \    c = *eptr; \
592    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; \  
       } \  
     }  
593    
594  /* 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
595  pointer. */  pointer. */
596    
597  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
598    c = *eptr; \    c = *eptr; \
599    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8(c, eptr);
600    
601    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
602    the pointer. */
603    
604    #define GETUTF8INC(c, eptr) \
605      { \      { \
606      int gcii; \      if ((c & 0x20) == 0) \
607      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
608      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
609      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
610      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
611          eptr += 2; \
612          } \
613        else if ((c & 0x08) == 0) \
614          { \
615          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
616              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
617          eptr += 3; \
618          } \
619        else if ((c & 0x04) == 0) \
620          { \
621          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
622              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
623              (eptr[3] & 0x3f); \
624          eptr += 4; \
625          } \
626        else \
627        { \        { \
628        gcss -= 6; \        c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
629        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
630              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
631          eptr += 5; \
632        } \        } \
633      }      }
634    
# Line 460  know we are in UTF-8 mode. */ Line 637  know we are in UTF-8 mode. */
637    
638  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
639    c = *eptr++; \    c = *eptr++; \
640    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; \  
       } \  
     }  
641    
642  /* 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.
643    This is called when we don't know if we are in UTF-8 mode. */
644    
645  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
646    c = *eptr++; \    c = *eptr++; \
647    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
648    
649    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
650    advancing the pointer, incrementing the length. */
651    
652    #define GETUTF8LEN(c, eptr, len) \
653      { \      { \
654      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
655      int gcss = 6*gcaa; \        { \
656      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
657      while (gcaa-- > 0) \        len++; \
658          } \
659        else if ((c & 0x10)  == 0) \
660        { \        { \
661        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
662        c |= (*eptr++ & 0x3f) << gcss; \        len += 2; \
663          } \
664        else if ((c & 0x08)  == 0) \
665          {\
666          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
667              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
668          len += 3; \
669          } \
670        else if ((c & 0x04)  == 0) \
671          { \
672          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
673              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
674              (eptr[4] & 0x3f); \
675          len += 4; \
676          } \
677        else \
678          {\
679          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
680              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
681              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
682          len += 5; \
683        } \        } \
684      }      }
685    
# Line 493  if there are extra bytes. This is called Line 688  if there are extra bytes. This is called
688    
689  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
690    c = *eptr; \    c = *eptr; \
691    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; \  
     }  
692    
693  /* 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
694  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
695  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
696    
697  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
698    c = *eptr; \    c = *eptr; \
699    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; \  
     }  
700    
701  /* 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
702  it is. This is called only in UTF-8 mode - we don't put a test within the macro  it is. This is called only in UTF-8 mode - we don't put a test within the macro
# Line 533  because almost all calls are already wit Line 704  because almost all calls are already wit
704    
705  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
706    
707    /* Same as above, just in the other direction. */
708    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
709    
710    /* Same as above, but it allows a fully customizable form. */
711    #define INTERNALCHAR(condition, eptr, action) \
712      while((condition) && ((eptr) & 0xc0) == 0x80) action
713    
714    #else /* COMPILE_PCRE8 */
715    
716    #ifdef COMPILE_PCRE16
717    
718    /* Tests whether the code point needs extra characters to decode. */
719    
720    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
721    
722    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
723    Otherwise it has an undefined behaviour. */
724    
725    #define GET_EXTRALEN(c) 1
726    
727    /* Base macro to pick up the low surrogate of a UTF-16 character, not
728    advancing the pointer. */
729    
730    #define GETUTF16(c, eptr) \
731       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
732    
733    /* Get the next UTF-16 character, not advancing the pointer. This is called when
734    we know we are in UTF-16 mode. */
735    
736    #define GETCHAR(c, eptr) \
737      c = *eptr; \
738      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
739    
740    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
741    pointer. */
742    
743    #define GETCHARTEST(c, eptr) \
744      c = *eptr; \
745      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
746    
747    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
748    the pointer. */
749    
750    #define GETUTF16INC(c, eptr) \
751       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; eptr++; }
752    
753    /* Get the next UTF-16 character, advancing the pointer. This is called when we
754    know we are in UTF-16 mode. */
755    
756    #define GETCHARINC(c, eptr) \
757      c = *eptr++; \
758      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
759    
760    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
761    This is called when we don't know if we are in UTF-16 mode. */
762    
763    #define GETCHARINCTEST(c, eptr) \
764      c = *eptr++; \
765      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
766    
767    /* Base macro to pick up the low surrogate of a UTF-16 character, not
768    advancing the pointer, incrementing the length. */
769    
770    #define GETUTF16LEN(c, eptr, len) \
771       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
772    
773    /* Get the next UTF-16 character, not advancing the pointer, incrementing
774    length if there is a low surrogate. This is called when we know we are in
775    UTF-16 mode. */
776    
777    #define GETCHARLEN(c, eptr, len) \
778      c = *eptr; \
779      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
780    
781    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
782    pointer, incrementing length if there is a low surrogate. This is called when
783    we do not know if we are in UTF-16 mode. */
784    
785    #define GETCHARLENTEST(c, eptr, len) \
786      c = *eptr; \
787      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
788    
789    /* If the pointer is not at the start of a character, move it back until
790    it is. This is called only in UTF-16 mode - we don't put a test within the
791    macro because almost all calls are already within a block of UTF-16 only
792    code. */
793    
794    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
795    
796    /* Same as above, just in the other direction. */
797    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
798    
799    /* Same as above, but it allows a fully customizable form. */
800    #define INTERNALCHAR(condition, eptr, action) \
801      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
802    
803  #endif  #endif
804    
805    #endif /* COMPILE_PCRE8 */
806    
807    #endif  /* SUPPORT_UTF */
808    
809    
810  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
811  Standard C system should have one. */  Standard C system should have one. */
# Line 544  Standard C system should have one. */ Line 815  Standard C system should have one. */
815  #endif  #endif
816    
817    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
818  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
819  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
820  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
# Line 555  the restrictions on partial matching hav Line 822  the restrictions on partial matching hav
822  compatibility. */  compatibility. */
823    
824  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
825  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_char is set */
826  #define PCRE_REQCHSET      0x0004  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
827  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
828  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
829  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
830    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
831    #define PCRE_FCH_CASELESS  0x0080  /* caseless first char */
832    #define PCRE_RCH_CASELESS  0x0100  /* caseless requested char */
833    
834  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
835    
836  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
837  #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */  #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
838    
839  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
840  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 577  time, run time, or study time, respectiv Line 847  time, run time, or study time, respectiv
847     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
848     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| \
849     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
850     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
851    
852  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
853    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
# Line 590  time, run time, or study time, respectiv Line 860  time, run time, or study time, respectiv
860     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
861     PCRE_NO_START_OPTIMIZE)     PCRE_NO_START_OPTIMIZE)
862    
863  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
864       PCRE_STUDY_JIT_COMPILE
865    
866  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. Also used
867  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 607  req_byte match. */ Line 878  req_byte match. */
878    
879  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
880    
 /* 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 */  
   
881  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
882  environments where these macros are defined elsewhere. Unfortunately, there  environments where these macros are defined elsewhere. Unfortunately, there
883  is no way to do the same for the typedef. */  is no way to do the same for the typedef. */
# Line 872  so that PCRE works on both ASCII and EBC Line 1137  so that PCRE works on both ASCII and EBC
1137  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
1138  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
1139  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
1140    #define STRING_MARK0                "MARK\0"
1141  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
1142  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
1143  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 893  so that PCRE works on both ASCII and EBC Line 1159  so that PCRE works on both ASCII and EBC
1159    
1160  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
1161    
1162  #define STRING_CR_RIGHTPAR          "CR)"  #define STRING_CR_RIGHTPAR             "CR)"
1163  #define STRING_LF_RIGHTPAR          "LF)"  #define STRING_LF_RIGHTPAR             "LF)"
1164  #define STRING_CRLF_RIGHTPAR        "CRLF)"  #define STRING_CRLF_RIGHTPAR           "CRLF)"
1165  #define STRING_ANY_RIGHTPAR         "ANY)"  #define STRING_ANY_RIGHTPAR            "ANY)"
1166  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1167  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1168  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1169  #define STRING_UTF8_RIGHTPAR        "UTF8)"  #define STRING_UTF8_RIGHTPAR           "UTF8)"
1170    #define STRING_UCP_RIGHTPAR            "UCP)"
1171    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1172    
1173  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF8 */
1174    
# Line 1124  only. */ Line 1392  only. */
1392  #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"
1393  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1394  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1395    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1396  #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"
1397  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1398  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1145  only. */ Line 1414  only. */
1414    
1415  #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
1416    
1417  #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1418  #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1419  #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
1420  #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
1421  #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
1422  #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
1423  #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
1424  #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1425    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1426    #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
1427    
1428  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
1429    
# Line 1185  only. */ Line 1456  only. */
1456    
1457  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1458  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1459  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1460  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1461  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1462    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1463    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1464    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1465    #define PT_WORD       8    /* Word - L plus N plus underscore */
1466    
1467  /* 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
1468  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1469    
1470  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1471  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 1204  contain UTF-8 characters with values gre Line 1479  contain UTF-8 characters with values gre
1479  /* 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
1480  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
1481  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
1482  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
1483  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
1484  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1485    non-DOTALL mode, "." behaves like \N.
1486    
1487    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1488    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1489    They must be contiguous, and remain in order so that the replacements can be
1490    looked up from a table.
1491    
1492  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
1493  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
# Line 1216  put in between that don't consume a char Line 1497  put in between that don't consume a char
1497  */  */
1498    
1499  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,
1500         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,
1501         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,
1502           ESC_E, ESC_Q, ESC_g, ESC_k,
1503           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1504         ESC_REF };         ESC_REF };
1505    
   
1506  /* 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
1507  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.
1508    
# Line 1244  enum { Line 1526  enum {
1526    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1527    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1528    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1529    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1530    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1531      OP_ALLANY,         /* 13 Match any character */
1532    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 */
1533    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1534    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1258  enum { Line 1541  enum {
1541    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. */
1542    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1543    
1544    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1545    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1546    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1547    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1548    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1549    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1550      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1551    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1552    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1553    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1554    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1555    OP_QUERY,          /* 35 */  
1556    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1557    
1558    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1559    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1560    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1561      OP_MINPLUS,        /* 36 */
1562    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1563    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1564    OP_POSQUERY,       /* 42 Posesssified query */  
1565    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1566      OP_MINUPTO,        /* 40 */
1567    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1568    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1569    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1570    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1571    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1572    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1573    
1574    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1575    OP_NOTMINUPTO,     /* 51 */  
1576    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1577      OP_MINSTARI,       /* 47 */
1578    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1579    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1580    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1581    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1582    
1583    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1584    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1585    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1586    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1587    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1588    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1589      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1590    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1591    OP_TYPEMINUPTO,    /* 64 */  
1592    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1593      /**** Negated single character, caseful; must precede the caseless ones ****/
1594    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1595    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1596    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1597    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1598      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1599    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1600    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1601    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1602    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1603    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1604    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1605    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1606    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1607      OP_NOTPOSPLUS,     /* 69 */
1608    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1609    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1610                             class - the difference is relevant only when a UTF-8  
1611                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1612    
1613    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1614                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1615      OP_NOTPLUSI,       /* 74 */
1616    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1617    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1618    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1619    
1620    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1621    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1622    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1623    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1624      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1625    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1626      OP_NOTPOSQUERYI,   /* 83 */
1627    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1628    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1629    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1630    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1631    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1632      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1633    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1634    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. */
1635      OP_TYPEQUERY,      /* 89 */
1636    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1637    OP_BRA,            /* 94 Start of non-capturing bracket */  
1638    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1639    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1640      OP_TYPEEXACT,      /* 93 Exactly n matches */
1641    
1642      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1643      OP_TYPEPOSPLUS,    /* 95 */
1644      OP_TYPEPOSQUERY,   /* 96 */
1645      OP_TYPEPOSUPTO,    /* 97 */
1646    
1647      /* These are used for character classes and back references; only the
1648      first six are the same as the sets above. */
1649    
1650      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1651      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1652      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1653      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1654      OP_CRQUERY,        /* 102 */
1655      OP_CRMINQUERY,     /* 103 */
1656    
1657      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1658      OP_CRMINRANGE,     /* 105 */
1659    
1660      /* End of quantifier opcodes */
1661    
1662      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1663      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1664                                  class - the difference is relevant only when a
1665                                  character > 255 is encountered. */
1666      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1667                                  class. This does both positive and negative. */
1668      OP_REF,            /* 109 Match a back reference, casefully */
1669      OP_REFI,           /* 110 Match a back reference, caselessly */
1670      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1671      OP_CALLOUT,        /* 112 Call out to external function if provided */
1672    
1673      OP_ALT,            /* 113 Start of alternation */
1674      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1675      OP_KETRMAX,        /* 115 These two must remain together and in this */
1676      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1677      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1678    
1679      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1680      asserts must remain in order. */
1681    
1682      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1683      OP_ASSERT,         /* 119 Positive lookahead */
1684      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1685      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1686      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1687    
1688      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1689      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1690      subpattern that isn't an assertion. The POS versions must immediately follow
1691      the non-POS versions in each case. */
1692    
1693      OP_ONCE,           /* 123 Atomic group, contains captures */
1694      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1695      OP_BRA,            /* 125 Start of non-capturing bracket */
1696      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1697      OP_CBRA,           /* 127 Start of capturing bracket */
1698      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1699      OP_COND,           /* 129 Conditional group */
1700    
1701    /* 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
1702    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1703    
1704    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1705    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1706    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1707      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1708      OP_SCOND,          /* 134 Conditional group, check empty */
1709    
1710    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
1711    
1712    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 135 Used to hold a capture number as condition */
1713    OP_NCREF,          /* 101 Same, but generaged by a name reference*/    OP_NCREF,          /* 136 Same, but generated by a name reference*/
1714    OP_RREF,           /* 102 Used to hold a recursion number as condition */    OP_RREF,           /* 137 Used to hold a recursion number as condition */
1715    OP_NRREF,          /* 103 Same, but generaged by a name reference*/    OP_NRREF,          /* 138 Same, but generated by a name reference*/
1716    OP_DEF,            /* 104 The DEFINE condition */    OP_DEF,            /* 139 The DEFINE condition */
1717    
1718    OP_BRAZERO,        /* 105 These two must remain together and in this */    OP_BRAZERO,        /* 140 These two must remain together and in this */
1719    OP_BRAMINZERO,     /* 106 order. */    OP_BRAMINZERO,     /* 141 order. */
1720      OP_BRAPOSZERO,     /* 142 */
1721    
1722    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1723    
1724    OP_PRUNE,          /* 107 */    OP_MARK,           /* 143 always has an argument */
1725    OP_SKIP,           /* 108 */    OP_PRUNE,          /* 144 */
1726    OP_THEN,           /* 109 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1727    OP_COMMIT,         /* 110 */    OP_SKIP,           /* 146 */
1728      OP_SKIP_ARG,       /* 147 same, but with argument */
1729      OP_THEN,           /* 148 */
1730      OP_THEN_ARG,       /* 149 same, but with argument */
1731      OP_COMMIT,         /* 150 */
1732    
1733    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1734    
1735    OP_FAIL,           /* 111 */    OP_FAIL,           /* 151 */
1736    OP_ACCEPT,         /* 112 */    OP_ACCEPT,         /* 152 */
1737    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1738      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1739    
1740    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1741    
1742    OP_SKIPZERO        /* 114 */    OP_SKIPZERO,       /* 155 */
1743    
1744      /* This is not an opcode, but is used to check that tables indexed by opcode
1745      are the correct length, in order to catch updating errors - there have been
1746      some in the past. */
1747    
1748      OP_TABLE_LENGTH
1749  };  };
1750    
1751  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1752  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
1753  called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */  called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1754    
1755    
1756  /* 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
1757  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
1758    only in pcre_printint.c, which fills out the full names in many cases (and in
1759    some cases doesn't actually use these names at all). */
1760    
1761  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1762    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1763    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1764    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1765    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1766    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1767    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1768      "{", "{", "{",                                                  \
1769    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1770    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1771      "{", "{", "{",                                                  \
1772      "*+","++", "?+", "{",                                           \
1773      "*", "*?", "+", "+?", "?", "??",                                \
1774      "{", "{", "{",                                                  \
1775      "*+","++", "?+", "{",                                           \
1776      "*", "*?", "+", "+?", "?", "??",                                \
1777      "{", "{", "{",                                                  \
1778    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1779    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1780    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1781    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1782    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1783    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1784    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1785    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1786      "Once", "Once_NC",                                              \
1787      "Bra", "BraPos", "CBra", "CBraPos",                             \
1788      "Cond",                                                         \
1789      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1790      "SCond",                                                        \
1791    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1792    "Brazero", "Braminzero",                                        \    "Brazero", "Braminzero", "Braposzero",                          \
1793    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1794      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1795      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1796    "Close", "Skip zero"    "Close", "Skip zero"
1797    
1798    
# Line 1436  in UTF-8 mode. The code that uses this t Line 1810  in UTF-8 mode. The code that uses this t
1810    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1811    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1812    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1813    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1814    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1815    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1816      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1817    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1818    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1819    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1820    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1821    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1822    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1823    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1824      2+IMM2_SIZE,                   /* exact                                  */ \
1825      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1826      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1827      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1828      2+IMM2_SIZE,                   /* exact I                                */ \
1829      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1830    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1831    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1832    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1833    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1834      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1835      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1836      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1837      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1838      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1839    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1840    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1841    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1842    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
1843      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1844    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1845    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1846    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1847   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
1848   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
1849    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1850    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1851      1+IMM2_SIZE,                   /* REFI                                   */ \
1852    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1853    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1854    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1855    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1856    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1857    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1858      1+LINK_SIZE,                   /* KetRpos                                */ \
1859      1+LINK_SIZE,                   /* Reverse                                */ \
1860    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1861    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1862    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1863    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1864    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1865      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1866    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1867    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
1868      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1869      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1870    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1871    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1872    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1873      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1874      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1875    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1876    3, 3,                          /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1877    3, 3,                          /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1878    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1879    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1880    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1881    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1882      1, 3,                          /* THEN, THEN_ARG                         */ \
1883      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1884      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1885    
1886  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1887  condition. */  condition. */
1888    
1889  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1890    
1891  /* 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
1892  tracked. */  easily be tracked. When a new number is added, the table called eint in
1893    pcreposix.c must be updated. */
1894    
1895  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1896         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1501  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1898  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1898         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1899         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1900         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1901         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1902           ERR70, ERR71, ERRCOUNT };
1903    
1904  /* 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
1905  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 1528  typedef struct real_pcre { Line 1926  typedef struct real_pcre {
1926    pcre_uint16 dummy1;             /* For future use */    pcre_uint16 dummy1;             /* For future use */
1927    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1928    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1929    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
1930    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
1931    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
1932    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
1933    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1934    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1935    
1936    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1937    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1938  } real_pcre;  } real_pcre;
1939    
1940  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
# Line 1545  remark (see NOTE above) about extending Line 1943  remark (see NOTE above) about extending
1943  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1944    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1945    pcre_uint32 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
1946    uschar start_bits[32];          /* Starting char bits */    pcre_uint8 start_bits[32];      /* Starting char bits */
1947    pcre_uint32 minlength;          /* Minimum subject length */    pcre_uint32 minlength;          /* Minimum subject length */
1948  } pcre_study_data;  } pcre_study_data;
1949    
1950  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1951  compiling, so that instructions to close them can be compiled when (*ACCEPT) is  compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1952  encountered. */  encountered. This is also used to identify subpatterns that contain recursive
1953    back references to themselves, so that they can be made atomic. */
1954    
1955  typedef struct open_capitem {  typedef struct open_capitem {
1956    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1957    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1958      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1959  } open_capitem;  } open_capitem;
1960    
1961  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1962  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1963    
1964  typedef struct compile_data {  typedef struct compile_data {
1965    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1966    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1967    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1968    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1969    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1970    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1971    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1972    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1973    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;          /* Chain of open capture items */
1974    uschar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
1975    uschar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
1976    int  names_found;             /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
1977    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
1978    int  bracount;                /* Count of capturing parens as we compile */    int  bracount;                    /* Count of capturing parens as we compile */
1979    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
1980    int  top_backref;             /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
1981    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
1982    int  external_options;        /* External (initial) options */    int  assert_depth;                /* Depth of nested assertions */
1983    int  external_flags;          /* External flag bits to be set */    int  external_options;            /* External (initial) options */
1984    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_flags;              /* External flag bits to be set */
1985    BOOL had_accept;              /* (*ACCEPT) encountered */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1986    BOOL check_lookbehind;        /* Lookbehinds need later checking */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1987    int  nltype;                  /* Newline type */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1988    int  nllen;                   /* Newline string length */    int  nltype;                      /* Newline type */
1989    uschar nl[4];                 /* Newline string when fixed length */    int  nllen;                       /* Newline string length */
1990      pcre_uchar nl[4];                 /* Newline string when fixed length */
1991  } compile_data;  } compile_data;
1992    
1993  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1994  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1995    
1996  typedef struct branch_chain {  typedef struct branch_chain {
1997    struct branch_chain *outer;    struct branch_chain *outer;
1998    uschar *current_branch;    pcre_uchar *current_branch;
1999  } branch_chain;  } branch_chain;
2000    
2001  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2002  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2003    
2004  typedef struct recursion_info {  typedef struct recursion_info {
2005    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2006    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2007    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2008    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
2009    int *offset_save;             /* Pointer to start of saved offsets */    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
   int save_offset_top;          /* Current value of offset_top */  
2010  } recursion_info;  } recursion_info;
2011    
2012    /* A similar structure for pcre_dfa_exec(). */
2013    
2014    typedef struct dfa_recursion_info {
2015      struct dfa_recursion_info *prevrec;
2016      int group_num;
2017      PCRE_PUCHAR subject_position;
2018    } dfa_recursion_info;
2019    
2020  /* 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
2021  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
2022  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
2023    pcre_exec(). */
2024    
2025  typedef struct eptrblock {  typedef struct eptrblock {
2026    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2027    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2028  } eptrblock;  } eptrblock;
2029    
2030    
# Line 1627  typedef struct match_data { Line 2035  typedef struct match_data {
2035    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2036    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2037    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2038    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2039    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2040    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2041    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2042    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2043    int    name_count;            /* Number of names in name table */    int    name_count;              /* Number of names in name table */
2044    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
2045    uschar *name_table;           /* Table of names */    pcre_uchar *name_table;         /* Table of names */
2046    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];               /* Newline string when fixed */
2047    const uschar *lcc;            /* Points to lower casing table */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2048    const uschar *ctypes;         /* Points to table of type maps */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2049    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;         /* Set if too many extractions */
2050    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                  /* NOTBOL flag */
2051    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                  /* NOTEOL flag */
2052    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2053    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2054    BOOL   endonly;               /* Dollar not before final \n */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2055    BOOL   notempty;              /* Empty string match not wanted */    BOOL   endonly;                 /* Dollar not before final \n */
2056    BOOL   notempty_atstart;      /* Empty string match at start not wanted */    BOOL   notempty;                /* Empty string match not wanted */
2057    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2058    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2059    const uschar *start_code;     /* For use when recursing */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2060    USPTR  start_subject;         /* Start of the subject string */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2061    USPTR  end_subject;           /* End of the subject string */    const  pcre_uchar *start_code;  /* For use when recursing */
2062    USPTR  start_match_ptr;       /* Start of matched string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2063    USPTR  end_match_ptr;         /* Subject position at end match */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2064    USPTR  start_used_ptr;        /* Earliest consulted character */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2065    int    partial;               /* PARTIAL options */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2066    int    end_offset_top;        /* Highwater mark at end of match */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2067    int    capture_last;          /* Most recent capture number */    int    partial;                 /* PARTIAL options */
2068    int    start_offset;          /* The start offset value */    int    end_offset_top;          /* Highwater mark at end of match */
2069    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    int    capture_last;            /* Most recent capture number */
2070    int    eptrn;                 /* Next free eptrblock */    int    start_offset;            /* The start offset value */
2071    recursion_info *recursive;    /* Linked list of recursion data */    int    match_function_type;     /* Set for certain special calls of MATCH() */
2072    void  *callout_data;          /* To pass back to callouts */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2073      int    eptrn;                   /* Next free eptrblock */
2074      recursion_info *recursive;      /* Linked list of recursion data */
2075      void  *callout_data;            /* To pass back to callouts */
2076      const  pcre_uchar *mark;        /* Mark pointer to pass back */
2077      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2078  } match_data;  } match_data;
2079    
2080  /* 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
2081  functions. */  functions. */
2082    
2083  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2084    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2085    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2086    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2087    const uschar *start_used_ptr; /* Earliest consulted character */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2088    const uschar *tables;         /* Character tables */    const pcre_uint8 *tables;         /* Character tables */
2089    int   start_offset;           /* The start offset value */    int   start_offset;               /* The start offset value */
2090    int   moptions;               /* Match options */    int   moptions;                   /* Match options */
2091    int   poptions;               /* Pattern options */    int   poptions;                   /* Pattern options */
2092    int    nltype;                /* Newline type */    int   nltype;                     /* Newline type */
2093    int    nllen;                 /* Newline string length */    int   nllen;                      /* Newline string length */
2094    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];                 /* Newline string when fixed */
2095    void  *callout_data;          /* To pass back to callouts */    void *callout_data;               /* To pass back to callouts */
2096      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2097  } dfa_match_data;  } dfa_match_data;
2098    
2099  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1715  total length. */ Line 2129  total length. */
2129  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2130  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2131    
2132    /* Internal function prefix */
2133    
2134    #ifdef COMPILE_PCRE8
2135    #define PRIV(name) _pcre_##name
2136    #else
2137    #ifdef COMPILE_PCRE16
2138    #define PRIV(name) _pcre16_##name
2139    #else
2140    #error Unsupported compiling mode
2141    #endif /* COMPILE_PCRE16 */
2142    #endif /* COMPILE_PCRE8 */
2143    
2144  /* 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
2145  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
2146  relocations in shared libraries, it now has an offset into a single string  relocations in shared libraries, it now has an offset into a single string
# Line 1732  of the exported public functions. They h Line 2158  of the exported public functions. They h
2158  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
2159  pcre_tables.c module. */  pcre_tables.c module. */
2160    
2161  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2162  extern const int    _pcre_utf8_table2[];  
2163  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1)[];
2164  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table1_size);
2165    extern const int            PRIV(utf8_table2)[];
2166    extern const int            PRIV(utf8_table3)[];
2167    extern const pcre_uint8     PRIV(utf8_table4)[];
2168    
2169  extern const int    _pcre_utf8_table1_size;  #endif /* COMPILE_PCRE8 */
2170    
2171  extern const char   _pcre_utt_names[];  extern const char           PRIV(utt_names)[];
2172  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table PRIV(utt)[];
2173  extern const int _pcre_utt_size;  extern const int            PRIV(utt_size);
2174    
2175  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2176    
2177  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2178    
2179    
2180  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2181  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
2182  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2183    
2184  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  /* String comparison functions. */
2185  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, 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(USPTR, int);  
 extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);  
 extern BOOL          _pcre_xclass(int, const uschar *);  
2186    
2187    #define STRCMP_UC_UC(str1, str2) \
2188      strcmp((char *)(str1), (char *)(str2))
2189    #define STRCMP_UC_C8(str1, str2) \
2190      strcmp((char *)(str1), (str2))
2191    #define STRNCMP_UC_UC(str1, str2, num) \
2192      strncmp((char *)(str1), (char *)(str2), (num))
2193    #define STRNCMP_UC_C8(str1, str2, num) \
2194      strncmp((char *)(str1), (str2), (num))
2195    #define STRLEN_UC(str) strlen((const char *)str)
2196    
2197    #else
2198    
2199    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2200                               const pcre_uchar *);
2201    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2202                               const char *);
2203    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2204                               const pcre_uchar *, unsigned int num);
2205    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2206                               const char *, unsigned int num);
2207    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2208    
2209    #define STRCMP_UC_UC(str1, str2) \
2210      PRIV(strcmp_uc_uc)((str1), (str2))
2211    #define STRCMP_UC_C8(str1, str2) \
2212      PRIV(strcmp_uc_c8)((str1), (str2))
2213    #define STRNCMP_UC_UC(str1, str2, num) \
2214      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2215    #define STRNCMP_UC_C8(str1, str2, num) \
2216      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2217    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2218    
2219    #endif /* COMPILE_PCRE8 */
2220    
2221    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2222    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2223                               int *, BOOL);
2224    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2225    extern real_pcre        *PRIV(try_flipped)(const real_pcre *, real_pcre *,
2226                               const pcre_study_data *, pcre_study_data *);
2227    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2228    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2229                               int *, BOOL);
2230    extern BOOL              PRIV(xclass)(int, const pcre_uchar *);
2231    
2232    #ifdef SUPPORT_JIT
2233    extern void              PRIV(jit_compile)(const real_pcre *, pcre_extra *);
2234    extern int               PRIV(jit_exec)(const real_pcre *, void *,
2235                               const pcre_uchar *, int, int, int, int, int *, int);
2236    extern void              PRIV(jit_free)(void *);
2237    #endif
2238    
2239  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
2240    
2241  typedef struct {  typedef struct {
2242    uschar script;    pcre_uint8 script;
2243    uschar chartype;    pcre_uint8 chartype;
2244    pcre_int32 other_case;    pcre_int32 other_case;
2245  } ucd_record;  } ucd_record;
2246    
2247  extern const ucd_record  _pcre_ucd_records[];  extern const ucd_record  PRIV(ucd_records)[];
2248  extern const uschar      _pcre_ucd_stage1[];  extern const pcre_uint8  PRIV(ucd_stage1)[];
2249  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint16 PRIV(ucd_stage2)[];
2250  extern const int         _pcre_ucp_gentype[];  extern const int         PRIV(ucp_gentype)[];
2251    #ifdef SUPPORT_JIT
2252    extern const int         PRIV(ucp_typerange)[];
2253    #endif
2254    
2255  /* UCD access macros */  /* UCD access macros */
2256    
2257  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
2258  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (PRIV(ucd_records) + \
2259          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2260          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2261    
2262  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2263  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2264  #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]  #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2265  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2266    
2267  #endif  #endif

Legend:
Removed from v.483  
changed lines
  Added in v.782

  ViewVC Help
Powered by ViewVC 1.1.5