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

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

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

code/trunk/pcre_internal.h revision 335 by ph10, Sat Apr 12 14:36:14 2008 UTC code/branches/pcre16/pcre_internal.h revision 789 by zherczeg, Wed Dec 7 14:36:26 2011 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2008 University of Cambridge             Copyright (c) 1997-2011 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 40  POSSIBILITY OF SUCH DAMAGE. Line 40  POSSIBILITY OF SUCH DAMAGE.
40    
41  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_" or "_pcre16_" depending on
44    the PRIV macro. */
45    
46  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
47  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
48    
49  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
50    
51  #if 0  #if 0
52  #define DEBUG  #define PCRE_DEBUG
53    #endif
54    
55    /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56    #ifndef COMPILE_PCRE16
57    #define COMPILE_PCRE8
58    #endif
59    
60    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
61    script prevents both being selected, but not everybody uses "configure". */
62    
63    #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
64    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
65    #endif
66    
67    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
68    "configure" script ensures this, but not everybody uses "configure". */
69    
70    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
71    #define SUPPORT_UTF8 1
72    #endif
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  #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
# Line 60  It turns out that the Mac Debugging.h he Line 95  It turns out that the Mac Debugging.h he
95  be absolutely sure we get our version. */  be absolutely sure we get our version. */
96    
97  #undef DPRINTF  #undef DPRINTF
98  #ifdef DEBUG  #ifdef PCRE_DEBUG
99  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
100  #else  #else
101  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
# Line 72  setjmp and stdarg are used is when NO_RE Line 107  setjmp and stdarg are used is when NO_RE
107    
108  #include <ctype.h>  #include <ctype.h>
109  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
110  #include <stddef.h>  #include <stddef.h>
111  #include <stdio.h>  #include <stdio.h>
112  #include <stdlib.h>  #include <stdlib.h>
# Line 132  PCRE_EXP_DATA_DEFN only if they are not Line 165  PCRE_EXP_DATA_DEFN only if they are not
165  #  endif  #  endif
166  #endif  #endif
167    
168  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* When compiling with the MSVC compiler, it is sometimes necessary to include
169    a "calling convention" before exported function names. (This is secondhand
170    information; I know nothing about MSVC myself). For example, something like
171    
172      void __cdecl function(....)
173    
174    might be needed. In order so make this easy, all the exported functions have
175    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
176    set, we ensure here that it has no effect. */
177    
178    #ifndef PCRE_CALL_CONVENTION
179    #define PCRE_CALL_CONVENTION
180    #endif
181    
182    /* 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;
193  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
194    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
195      typedef int pcre_int16;
196  #else  #else
197    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
198  #endif  #endif
199    
200  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
201    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
202      typedef int pcre_int32;
203  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
204    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
205      typedef long int pcre_int32;
206  #else  #else
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
211    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
213    arithmetic. Handle this by defining a macro for the appropriate type. If
214    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
215    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
219    #include <stdint.h>
220    #elif HAVE_INTTYPES_H
221    #include <inttypes.h>
222    #endif
223    
224    #if defined INT64_MAX || defined int64_t
225    #define INT64_OR_DOUBLE int64_t
226    #else
227    #define INT64_OR_DOUBLE double
228    #endif
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 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 unsigned char uschar;  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 184  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 198  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 217  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 293  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 329  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 346  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 353  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 NEXTCHAR(p) p++;  /* #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 ACROSSCHAR(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  #else   /* SUPPORT_UTF8 */  /* Tests whether the code point needs extra characters to decode. */
557    
558  /* Advance a character pointer one byte in non-UTF-8 mode and by one character  #define HAS_EXTRALEN(c) ((c) >= 0xc0)
 in UTF-8 mode. */  
559    
560  #define NEXTCHAR(p) \  /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
561    p++; \  Otherwise it has an undefined behaviour. */
562    if (utf8) { while((*p & 0xc0) == 0x80) p++; }  
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    #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        gcss -= 6; \        c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
616        c |= (eptr[gcii] & 0x3f) << gcss; \            ((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          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
629              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
630              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
631          eptr += 5; \
632        } \        } \
633      }      }
634    
# Line 421  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) \
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
655        { \        { \
656        gcss -= 6; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
657        c |= (*eptr++ & 0x3f) << gcss; \        len++; \
658          } \
659        else if ((c & 0x10)  == 0) \
660          { \
661          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
662          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 454  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);
692      { \  
693      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
694      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
695      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
696      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
697      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
698        { \    c = *eptr; \
699        gcss -= 6; \    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       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 474  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 ACROSSCHAR(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++ & 0x3ff)) + 0x10000; }
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 ACROSSCHAR(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 485  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. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
821    the restrictions on partial matching have been lifted. It remains for backwards
822    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  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 510  time, run time, or study time, respectiv Line 842  time, run time, or study time, respectiv
842  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
843                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
844    
845  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
846    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
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_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_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
854     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
855       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
856    
857  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
858    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
859     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
860     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
861       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 542  req_byte match. */ Line 878  req_byte match. */
878    
879  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
880    
881  /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
882  variable-length repeat, or a anything other than literal characters. */  environments where these macros are defined elsewhere. Unfortunately, there
883    is no way to do the same for the typedef. */
 #define REQ_CASELESS 0x0100    /* indicates caselessness */  
 #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  
   
 /* Miscellaneous definitions */  
884    
885  typedef int BOOL;  typedef int BOOL;
886    
887    #ifndef FALSE
888  #define FALSE   0  #define FALSE   0
889  #define TRUE    1  #define TRUE    1
890    #endif
891    
892    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
893    character constants like '*' because the compiler would emit their EBCDIC code,
894    which is different from their ASCII/UTF-8 code. Instead we define macros for
895    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
896    is enabled. When UTF-8 support is not enabled, the definitions use character
897    literals. Both character and string versions of each character are needed, and
898    there are some longer strings as well.
899    
900    This means that, on EBCDIC platforms, the PCRE library can handle either
901    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
902    would need different lookups depending on whether PCRE_UTF8 was set or not.
903    This would make it impossible to use characters in switch/case statements,
904    which would reduce performance. For a theoretical use (which nobody has asked
905    for) in a minority area (EBCDIC platforms), this is not sensible. Any
906    application that did need both could compile two versions of the library, using
907    macros to give the functions distinct names. */
908    
909    #ifndef SUPPORT_UTF8
910    
911    /* UTF-8 support is not enabled; use the platform-dependent character literals
912    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
913    
914    #define CHAR_HT                     '\t'
915    #define CHAR_VT                     '\v'
916    #define CHAR_FF                     '\f'
917    #define CHAR_CR                     '\r'
918    #define CHAR_NL                     '\n'
919    #define CHAR_BS                     '\b'
920    #define CHAR_BEL                    '\a'
921    #ifdef EBCDIC
922    #define CHAR_ESC                    '\047'
923    #define CHAR_DEL                    '\007'
924    #else
925    #define CHAR_ESC                    '\033'
926    #define CHAR_DEL                    '\177'
927    #endif
928    
929    #define CHAR_SPACE                  ' '
930    #define CHAR_EXCLAMATION_MARK       '!'
931    #define CHAR_QUOTATION_MARK         '"'
932    #define CHAR_NUMBER_SIGN            '#'
933    #define CHAR_DOLLAR_SIGN            '$'
934    #define CHAR_PERCENT_SIGN           '%'
935    #define CHAR_AMPERSAND              '&'
936    #define CHAR_APOSTROPHE             '\''
937    #define CHAR_LEFT_PARENTHESIS       '('
938    #define CHAR_RIGHT_PARENTHESIS      ')'
939    #define CHAR_ASTERISK               '*'
940    #define CHAR_PLUS                   '+'
941    #define CHAR_COMMA                  ','
942    #define CHAR_MINUS                  '-'
943    #define CHAR_DOT                    '.'
944    #define CHAR_SLASH                  '/'
945    #define CHAR_0                      '0'
946    #define CHAR_1                      '1'
947    #define CHAR_2                      '2'
948    #define CHAR_3                      '3'
949    #define CHAR_4                      '4'
950    #define CHAR_5                      '5'
951    #define CHAR_6                      '6'
952    #define CHAR_7                      '7'
953    #define CHAR_8                      '8'
954    #define CHAR_9                      '9'
955    #define CHAR_COLON                  ':'
956    #define CHAR_SEMICOLON              ';'
957    #define CHAR_LESS_THAN_SIGN         '<'
958    #define CHAR_EQUALS_SIGN            '='
959    #define CHAR_GREATER_THAN_SIGN      '>'
960    #define CHAR_QUESTION_MARK          '?'
961    #define CHAR_COMMERCIAL_AT          '@'
962    #define CHAR_A                      'A'
963    #define CHAR_B                      'B'
964    #define CHAR_C                      'C'
965    #define CHAR_D                      'D'
966    #define CHAR_E                      'E'
967    #define CHAR_F                      'F'
968    #define CHAR_G                      'G'
969    #define CHAR_H                      'H'
970    #define CHAR_I                      'I'
971    #define CHAR_J                      'J'
972    #define CHAR_K                      'K'
973    #define CHAR_L                      'L'
974    #define CHAR_M                      'M'
975    #define CHAR_N                      'N'
976    #define CHAR_O                      'O'
977    #define CHAR_P                      'P'
978    #define CHAR_Q                      'Q'
979    #define CHAR_R                      'R'
980    #define CHAR_S                      'S'
981    #define CHAR_T                      'T'
982    #define CHAR_U                      'U'
983    #define CHAR_V                      'V'
984    #define CHAR_W                      'W'
985    #define CHAR_X                      'X'
986    #define CHAR_Y                      'Y'
987    #define CHAR_Z                      'Z'
988    #define CHAR_LEFT_SQUARE_BRACKET    '['
989    #define CHAR_BACKSLASH              '\\'
990    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
991    #define CHAR_CIRCUMFLEX_ACCENT      '^'
992    #define CHAR_UNDERSCORE             '_'
993    #define CHAR_GRAVE_ACCENT           '`'
994    #define CHAR_a                      'a'
995    #define CHAR_b                      'b'
996    #define CHAR_c                      'c'
997    #define CHAR_d                      'd'
998    #define CHAR_e                      'e'
999    #define CHAR_f                      'f'
1000    #define CHAR_g                      'g'
1001    #define CHAR_h                      'h'
1002    #define CHAR_i                      'i'
1003    #define CHAR_j                      'j'
1004    #define CHAR_k                      'k'
1005    #define CHAR_l                      'l'
1006    #define CHAR_m                      'm'
1007    #define CHAR_n                      'n'
1008    #define CHAR_o                      'o'
1009    #define CHAR_p                      'p'
1010    #define CHAR_q                      'q'
1011    #define CHAR_r                      'r'
1012    #define CHAR_s                      's'
1013    #define CHAR_t                      't'
1014    #define CHAR_u                      'u'
1015    #define CHAR_v                      'v'
1016    #define CHAR_w                      'w'
1017    #define CHAR_x                      'x'
1018    #define CHAR_y                      'y'
1019    #define CHAR_z                      'z'
1020    #define CHAR_LEFT_CURLY_BRACKET     '{'
1021    #define CHAR_VERTICAL_LINE          '|'
1022    #define CHAR_RIGHT_CURLY_BRACKET    '}'
1023    #define CHAR_TILDE                  '~'
1024    
1025    #define STR_HT                      "\t"
1026    #define STR_VT                      "\v"
1027    #define STR_FF                      "\f"
1028    #define STR_CR                      "\r"
1029    #define STR_NL                      "\n"
1030    #define STR_BS                      "\b"
1031    #define STR_BEL                     "\a"
1032    #ifdef EBCDIC
1033    #define STR_ESC                     "\047"
1034    #define STR_DEL                     "\007"
1035    #else
1036    #define STR_ESC                     "\033"
1037    #define STR_DEL                     "\177"
1038    #endif
1039    
1040    #define STR_SPACE                   " "
1041    #define STR_EXCLAMATION_MARK        "!"
1042    #define STR_QUOTATION_MARK          "\""
1043    #define STR_NUMBER_SIGN             "#"
1044    #define STR_DOLLAR_SIGN             "$"
1045    #define STR_PERCENT_SIGN            "%"
1046    #define STR_AMPERSAND               "&"
1047    #define STR_APOSTROPHE              "'"
1048    #define STR_LEFT_PARENTHESIS        "("
1049    #define STR_RIGHT_PARENTHESIS       ")"
1050    #define STR_ASTERISK                "*"
1051    #define STR_PLUS                    "+"
1052    #define STR_COMMA                   ","
1053    #define STR_MINUS                   "-"
1054    #define STR_DOT                     "."
1055    #define STR_SLASH                   "/"
1056    #define STR_0                       "0"
1057    #define STR_1                       "1"
1058    #define STR_2                       "2"
1059    #define STR_3                       "3"
1060    #define STR_4                       "4"
1061    #define STR_5                       "5"
1062    #define STR_6                       "6"
1063    #define STR_7                       "7"
1064    #define STR_8                       "8"
1065    #define STR_9                       "9"
1066    #define STR_COLON                   ":"
1067    #define STR_SEMICOLON               ";"
1068    #define STR_LESS_THAN_SIGN          "<"
1069    #define STR_EQUALS_SIGN             "="
1070    #define STR_GREATER_THAN_SIGN       ">"
1071    #define STR_QUESTION_MARK           "?"
1072    #define STR_COMMERCIAL_AT           "@"
1073    #define STR_A                       "A"
1074    #define STR_B                       "B"
1075    #define STR_C                       "C"
1076    #define STR_D                       "D"
1077    #define STR_E                       "E"
1078    #define STR_F                       "F"
1079    #define STR_G                       "G"
1080    #define STR_H                       "H"
1081    #define STR_I                       "I"
1082    #define STR_J                       "J"
1083    #define STR_K                       "K"
1084    #define STR_L                       "L"
1085    #define STR_M                       "M"
1086    #define STR_N                       "N"
1087    #define STR_O                       "O"
1088    #define STR_P                       "P"
1089    #define STR_Q                       "Q"
1090    #define STR_R                       "R"
1091    #define STR_S                       "S"
1092    #define STR_T                       "T"
1093    #define STR_U                       "U"
1094    #define STR_V                       "V"
1095    #define STR_W                       "W"
1096    #define STR_X                       "X"
1097    #define STR_Y                       "Y"
1098    #define STR_Z                       "Z"
1099    #define STR_LEFT_SQUARE_BRACKET     "["
1100    #define STR_BACKSLASH               "\\"
1101    #define STR_RIGHT_SQUARE_BRACKET    "]"
1102    #define STR_CIRCUMFLEX_ACCENT       "^"
1103    #define STR_UNDERSCORE              "_"
1104    #define STR_GRAVE_ACCENT            "`"
1105    #define STR_a                       "a"
1106    #define STR_b                       "b"
1107    #define STR_c                       "c"
1108    #define STR_d                       "d"
1109    #define STR_e                       "e"
1110    #define STR_f                       "f"
1111    #define STR_g                       "g"
1112    #define STR_h                       "h"
1113    #define STR_i                       "i"
1114    #define STR_j                       "j"
1115    #define STR_k                       "k"
1116    #define STR_l                       "l"
1117    #define STR_m                       "m"
1118    #define STR_n                       "n"
1119    #define STR_o                       "o"
1120    #define STR_p                       "p"
1121    #define STR_q                       "q"
1122    #define STR_r                       "r"
1123    #define STR_s                       "s"
1124    #define STR_t                       "t"
1125    #define STR_u                       "u"
1126    #define STR_v                       "v"
1127    #define STR_w                       "w"
1128    #define STR_x                       "x"
1129    #define STR_y                       "y"
1130    #define STR_z                       "z"
1131    #define STR_LEFT_CURLY_BRACKET      "{"
1132    #define STR_VERTICAL_LINE           "|"
1133    #define STR_RIGHT_CURLY_BRACKET     "}"
1134    #define STR_TILDE                   "~"
1135    
1136    #define STRING_ACCEPT0              "ACCEPT\0"
1137    #define STRING_COMMIT0              "COMMIT\0"
1138    #define STRING_F0                   "F\0"
1139    #define STRING_FAIL0                "FAIL\0"
1140    #define STRING_MARK0                "MARK\0"
1141    #define STRING_PRUNE0               "PRUNE\0"
1142    #define STRING_SKIP0                "SKIP\0"
1143    #define STRING_THEN                 "THEN"
1144    
1145    #define STRING_alpha0               "alpha\0"
1146    #define STRING_lower0               "lower\0"
1147    #define STRING_upper0               "upper\0"
1148    #define STRING_alnum0               "alnum\0"
1149    #define STRING_ascii0               "ascii\0"
1150    #define STRING_blank0               "blank\0"
1151    #define STRING_cntrl0               "cntrl\0"
1152    #define STRING_digit0               "digit\0"
1153    #define STRING_graph0               "graph\0"
1154    #define STRING_print0               "print\0"
1155    #define STRING_punct0               "punct\0"
1156    #define STRING_space0               "space\0"
1157    #define STRING_word0                "word\0"
1158    #define STRING_xdigit               "xdigit"
1159    
1160    #define STRING_DEFINE               "DEFINE"
1161    
1162    #define STRING_CR_RIGHTPAR             "CR)"
1163    #define STRING_LF_RIGHTPAR             "LF)"
1164    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1165    #define STRING_ANY_RIGHTPAR            "ANY)"
1166    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1167    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1168    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1169    #ifdef COMPILE_PCRE8
1170    #define STRING_UTF_RIGHTPAR            "UTF8)"
1171    #endif
1172    #ifdef COMPILE_PCRE16
1173    #define STRING_UTF_RIGHTPAR            "UTF16)"
1174    #endif
1175    #define STRING_UCP_RIGHTPAR            "UCP)"
1176    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1177    
1178    #else  /* SUPPORT_UTF8 */
1179    
1180    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1181    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1182    only. */
1183    
1184    #define CHAR_HT                     '\011'
1185    #define CHAR_VT                     '\013'
1186    #define CHAR_FF                     '\014'
1187    #define CHAR_CR                     '\015'
1188    #define CHAR_NL                     '\012'
1189    #define CHAR_BS                     '\010'
1190    #define CHAR_BEL                    '\007'
1191    #define CHAR_ESC                    '\033'
1192    #define CHAR_DEL                    '\177'
1193    
1194    #define CHAR_SPACE                  '\040'
1195    #define CHAR_EXCLAMATION_MARK       '\041'
1196    #define CHAR_QUOTATION_MARK         '\042'
1197    #define CHAR_NUMBER_SIGN            '\043'
1198    #define CHAR_DOLLAR_SIGN            '\044'
1199    #define CHAR_PERCENT_SIGN           '\045'
1200    #define CHAR_AMPERSAND              '\046'
1201    #define CHAR_APOSTROPHE             '\047'
1202    #define CHAR_LEFT_PARENTHESIS       '\050'
1203    #define CHAR_RIGHT_PARENTHESIS      '\051'
1204    #define CHAR_ASTERISK               '\052'
1205    #define CHAR_PLUS                   '\053'
1206    #define CHAR_COMMA                  '\054'
1207    #define CHAR_MINUS                  '\055'
1208    #define CHAR_DOT                    '\056'
1209    #define CHAR_SLASH                  '\057'
1210    #define CHAR_0                      '\060'
1211    #define CHAR_1                      '\061'
1212    #define CHAR_2                      '\062'
1213    #define CHAR_3                      '\063'
1214    #define CHAR_4                      '\064'
1215    #define CHAR_5                      '\065'
1216    #define CHAR_6                      '\066'
1217    #define CHAR_7                      '\067'
1218    #define CHAR_8                      '\070'
1219    #define CHAR_9                      '\071'
1220    #define CHAR_COLON                  '\072'
1221    #define CHAR_SEMICOLON              '\073'
1222    #define CHAR_LESS_THAN_SIGN         '\074'
1223    #define CHAR_EQUALS_SIGN            '\075'
1224    #define CHAR_GREATER_THAN_SIGN      '\076'
1225    #define CHAR_QUESTION_MARK          '\077'
1226    #define CHAR_COMMERCIAL_AT          '\100'
1227    #define CHAR_A                      '\101'
1228    #define CHAR_B                      '\102'
1229    #define CHAR_C                      '\103'
1230    #define CHAR_D                      '\104'
1231    #define CHAR_E                      '\105'
1232    #define CHAR_F                      '\106'
1233    #define CHAR_G                      '\107'
1234    #define CHAR_H                      '\110'
1235    #define CHAR_I                      '\111'
1236    #define CHAR_J                      '\112'
1237    #define CHAR_K                      '\113'
1238    #define CHAR_L                      '\114'
1239    #define CHAR_M                      '\115'
1240    #define CHAR_N                      '\116'
1241    #define CHAR_O                      '\117'
1242    #define CHAR_P                      '\120'
1243    #define CHAR_Q                      '\121'
1244    #define CHAR_R                      '\122'
1245    #define CHAR_S                      '\123'
1246    #define CHAR_T                      '\124'
1247    #define CHAR_U                      '\125'
1248    #define CHAR_V                      '\126'
1249    #define CHAR_W                      '\127'
1250    #define CHAR_X                      '\130'
1251    #define CHAR_Y                      '\131'
1252    #define CHAR_Z                      '\132'
1253    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1254    #define CHAR_BACKSLASH              '\134'
1255    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1256    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1257    #define CHAR_UNDERSCORE             '\137'
1258    #define CHAR_GRAVE_ACCENT           '\140'
1259    #define CHAR_a                      '\141'
1260    #define CHAR_b                      '\142'
1261    #define CHAR_c                      '\143'
1262    #define CHAR_d                      '\144'
1263    #define CHAR_e                      '\145'
1264    #define CHAR_f                      '\146'
1265    #define CHAR_g                      '\147'
1266    #define CHAR_h                      '\150'
1267    #define CHAR_i                      '\151'
1268    #define CHAR_j                      '\152'
1269    #define CHAR_k                      '\153'
1270    #define CHAR_l                      '\154'
1271    #define CHAR_m                      '\155'
1272    #define CHAR_n                      '\156'
1273    #define CHAR_o                      '\157'
1274    #define CHAR_p                      '\160'
1275    #define CHAR_q                      '\161'
1276    #define CHAR_r                      '\162'
1277    #define CHAR_s                      '\163'
1278    #define CHAR_t                      '\164'
1279    #define CHAR_u                      '\165'
1280    #define CHAR_v                      '\166'
1281    #define CHAR_w                      '\167'
1282    #define CHAR_x                      '\170'
1283    #define CHAR_y                      '\171'
1284    #define CHAR_z                      '\172'
1285    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1286    #define CHAR_VERTICAL_LINE          '\174'
1287    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1288    #define CHAR_TILDE                  '\176'
1289    
1290    #define STR_HT                      "\011"
1291    #define STR_VT                      "\013"
1292    #define STR_FF                      "\014"
1293    #define STR_CR                      "\015"
1294    #define STR_NL                      "\012"
1295    #define STR_BS                      "\010"
1296    #define STR_BEL                     "\007"
1297    #define STR_ESC                     "\033"
1298    #define STR_DEL                     "\177"
1299    
1300    #define STR_SPACE                   "\040"
1301    #define STR_EXCLAMATION_MARK        "\041"
1302    #define STR_QUOTATION_MARK          "\042"
1303    #define STR_NUMBER_SIGN             "\043"
1304    #define STR_DOLLAR_SIGN             "\044"
1305    #define STR_PERCENT_SIGN            "\045"
1306    #define STR_AMPERSAND               "\046"
1307    #define STR_APOSTROPHE              "\047"
1308    #define STR_LEFT_PARENTHESIS        "\050"
1309    #define STR_RIGHT_PARENTHESIS       "\051"
1310    #define STR_ASTERISK                "\052"
1311    #define STR_PLUS                    "\053"
1312    #define STR_COMMA                   "\054"
1313    #define STR_MINUS                   "\055"
1314    #define STR_DOT                     "\056"
1315    #define STR_SLASH                   "\057"
1316    #define STR_0                       "\060"
1317    #define STR_1                       "\061"
1318    #define STR_2                       "\062"
1319    #define STR_3                       "\063"
1320    #define STR_4                       "\064"
1321    #define STR_5                       "\065"
1322    #define STR_6                       "\066"
1323    #define STR_7                       "\067"
1324    #define STR_8                       "\070"
1325    #define STR_9                       "\071"
1326    #define STR_COLON                   "\072"
1327    #define STR_SEMICOLON               "\073"
1328    #define STR_LESS_THAN_SIGN          "\074"
1329    #define STR_EQUALS_SIGN             "\075"
1330    #define STR_GREATER_THAN_SIGN       "\076"
1331    #define STR_QUESTION_MARK           "\077"
1332    #define STR_COMMERCIAL_AT           "\100"
1333    #define STR_A                       "\101"
1334    #define STR_B                       "\102"
1335    #define STR_C                       "\103"
1336    #define STR_D                       "\104"
1337    #define STR_E                       "\105"
1338    #define STR_F                       "\106"
1339    #define STR_G                       "\107"
1340    #define STR_H                       "\110"
1341    #define STR_I                       "\111"
1342    #define STR_J                       "\112"
1343    #define STR_K                       "\113"
1344    #define STR_L                       "\114"
1345    #define STR_M                       "\115"
1346    #define STR_N                       "\116"
1347    #define STR_O                       "\117"
1348    #define STR_P                       "\120"
1349    #define STR_Q                       "\121"
1350    #define STR_R                       "\122"
1351    #define STR_S                       "\123"
1352    #define STR_T                       "\124"
1353    #define STR_U                       "\125"
1354    #define STR_V                       "\126"
1355    #define STR_W                       "\127"
1356    #define STR_X                       "\130"
1357    #define STR_Y                       "\131"
1358    #define STR_Z                       "\132"
1359    #define STR_LEFT_SQUARE_BRACKET     "\133"
1360    #define STR_BACKSLASH               "\134"
1361    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1362    #define STR_CIRCUMFLEX_ACCENT       "\136"
1363    #define STR_UNDERSCORE              "\137"
1364    #define STR_GRAVE_ACCENT            "\140"
1365    #define STR_a                       "\141"
1366    #define STR_b                       "\142"
1367    #define STR_c                       "\143"
1368    #define STR_d                       "\144"
1369    #define STR_e                       "\145"
1370    #define STR_f                       "\146"
1371    #define STR_g                       "\147"
1372    #define STR_h                       "\150"
1373    #define STR_i                       "\151"
1374    #define STR_j                       "\152"
1375    #define STR_k                       "\153"
1376    #define STR_l                       "\154"
1377    #define STR_m                       "\155"
1378    #define STR_n                       "\156"
1379    #define STR_o                       "\157"
1380    #define STR_p                       "\160"
1381    #define STR_q                       "\161"
1382    #define STR_r                       "\162"
1383    #define STR_s                       "\163"
1384    #define STR_t                       "\164"
1385    #define STR_u                       "\165"
1386    #define STR_v                       "\166"
1387    #define STR_w                       "\167"
1388    #define STR_x                       "\170"
1389    #define STR_y                       "\171"
1390    #define STR_z                       "\172"
1391    #define STR_LEFT_CURLY_BRACKET      "\173"
1392    #define STR_VERTICAL_LINE           "\174"
1393    #define STR_RIGHT_CURLY_BRACKET     "\175"
1394    #define STR_TILDE                   "\176"
1395    
1396    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1397    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1398    #define STRING_F0                   STR_F "\0"
1399    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1400    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1401    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1402    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1403    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1404    
1405    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1406    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1407    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1408    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1409    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1410    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1411    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1412    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1413    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1414    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1415    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1416    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1417    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1418    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1419    
1420    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1421    
1422    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1423    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1424    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1425    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1426    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1427    #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
1428    #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
1429    #ifdef COMPILE_PCRE8
1430    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1431    #endif
1432    #ifdef COMPILE_PCRE16
1433    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1434    #endif
1435    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1436    #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
1437    
1438    #endif  /* SUPPORT_UTF8 */
1439    
1440  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1441    
1442  #ifndef ESC_e  #ifndef ESC_e
1443  #define ESC_e 27  #define ESC_e CHAR_ESC
1444  #endif  #endif
1445    
1446  #ifndef ESC_f  #ifndef ESC_f
1447  #define ESC_f '\f'  #define ESC_f CHAR_FF
1448  #endif  #endif
1449    
1450  #ifndef ESC_n  #ifndef ESC_n
1451  #define ESC_n '\n'  #define ESC_n CHAR_NL
1452  #endif  #endif
1453    
1454  #ifndef ESC_r  #ifndef ESC_r
1455  #define ESC_r '\r'  #define ESC_r CHAR_CR
1456  #endif  #endif
1457    
1458  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1459  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1460    
1461  #ifndef ESC_tee  #ifndef ESC_tee
1462  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1463  #endif  #endif
1464    
1465  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1466    
1467  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1468  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1469  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1470  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1471  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1472    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1473    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1474    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1475    #define PT_WORD       8    /* Word - L plus N plus underscore */
1476    
1477  /* 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
1478  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1479    
1480  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1481  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 603  contain UTF-8 characters with values gre Line 1489  contain UTF-8 characters with values gre
1489  /* 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
1490  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
1491  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
1492  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
1493  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1494  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1495  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1496  ESC_Z to detect the types that may be repeated. These are the types that  
1497  consume characters. If any new escapes are put in between that don't consume a  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1498  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1499    They must be contiguous, and remain in order so that the replacements can be
1500    looked up from a table.
1501    
1502    The final escape must be ESC_REF as subsequent values are used for
1503    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1504    greater than ESC_b and less than ESC_Z to detect the types that may be
1505    repeated. These are the types that consume characters. If any new escapes are
1506    put in between that don't consume a character, that code will have to change.
1507    */
1508    
1509  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,
1510         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1511         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1512           ESC_E, ESC_Q, ESC_g, ESC_k,
1513           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1514           ESC_REF };
1515    
1516  /* 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
1517  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.
1518    
1519  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1520  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There are also tables called
1521  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1522    
1523  enum {  enum {
1524    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 639  enum { Line 1536  enum {
1536    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1537    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1538    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
   OP_ANY,            /* 12 Match any character */  
   OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */  
   OP_NOTPROP,        /* 14 \P (not Unicode property) */  
   OP_PROP,           /* 15 \p (Unicode property) */  
   OP_ANYNL,          /* 16 \R (any newline sequence) */  
   OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */  
   OP_HSPACE,         /* 18 \h (horizontal whitespace) */  
   OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */  
   OP_VSPACE,         /* 20 \v (vertical whitespace) */  
   OP_EXTUNI,         /* 21 \X (extended Unicode sequence */  
   OP_EODN,           /* 22 End of data or \n at end of data: \Z. */  
   OP_EOD,            /* 23 End of data: \z */  
   
   OP_OPT,            /* 24 Set runtime options */  
   OP_CIRC,           /* 25 Start of line - varies with multiline switch */  
   OP_DOLL,           /* 26 End of line - varies with multiline switch */  
   OP_CHAR,           /* 27 Match one character, casefully */  
   OP_CHARNC,         /* 28 Match one character, caselessly */  
   OP_NOT,            /* 29 Match one character, not the following one */  
   
   OP_STAR,           /* 30 The maximizing and minimizing versions of */  
   OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */  
   OP_PLUS,           /* 32 the minimizing one second. */  
   OP_MINPLUS,        /* 33 This first set applies to single characters.*/  
   OP_QUERY,          /* 34 */  
   OP_MINQUERY,       /* 35 */  
   
   OP_UPTO,           /* 36 From 0 to n matches */  
   OP_MINUPTO,        /* 37 */  
   OP_EXACT,          /* 38 Exactly n matches */  
   
   OP_POSSTAR,        /* 39 Possessified star */  
   OP_POSPLUS,        /* 40 Possessified plus */  
   OP_POSQUERY,       /* 41 Posesssified query */  
   OP_POSUPTO,        /* 42 Possessified upto */  
   
   OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
   OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */  
   OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */  
   OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */  
   OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */  
   OP_NOTMINQUERY,    /* 48 */  
   
   OP_NOTUPTO,        /* 49 From 0 to n matches */  
   OP_NOTMINUPTO,     /* 50 */  
   OP_NOTEXACT,       /* 51 Exactly n matches */  
   
   OP_NOTPOSSTAR,     /* 52 Possessified versions */  
   OP_NOTPOSPLUS,     /* 53 */  
   OP_NOTPOSQUERY,    /* 54 */  
   OP_NOTPOSUPTO,     /* 55 */  
   
   OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
   OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */  
   OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */  
   OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */  
   OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */  
   OP_TYPEMINQUERY,   /* 61 */  
   
   OP_TYPEUPTO,       /* 62 From 0 to n matches */  
   OP_TYPEMINUPTO,    /* 63 */  
   OP_TYPEEXACT,      /* 64 Exactly n matches */  
   
   OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
   OP_TYPEPOSPLUS,    /* 66 */  
   OP_TYPEPOSQUERY,   /* 67 */  
   OP_TYPEPOSUPTO,    /* 68 */  
   
   OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
   OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */  
   OP_CRPLUS,         /* 71 the minimizing one second. These codes must */  
   OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */  
   OP_CRQUERY,        /* 73 These are for character classes and back refs */  
   OP_CRMINQUERY,     /* 74 */  
   OP_CRRANGE,        /* 75 These are different to the three sets above. */  
   OP_CRMINRANGE,     /* 76 */  
   
   OP_CLASS,          /* 77 Match a character class, chars < 256 only */  
   OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative  
                            class - the difference is relevant only when a UTF-8  
                            character > 255 is encountered. */  
   
   OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the  
                            class. This does both positive and negative. */  
   
   OP_REF,            /* 80 Match a back reference */  
   OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */  
   OP_CALLOUT,        /* 82 Call out to external function if provided */  
   
   OP_ALT,            /* 83 Start of alternation */  
   OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */  
   OP_KETRMAX,        /* 85 These two must remain together and in this */  
   OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */  
   
   /* The assertions must come before BRA, CBRA, ONCE, and COND.*/  
   
   OP_ASSERT,         /* 87 Positive lookahead */  
   OP_ASSERT_NOT,     /* 88 Negative lookahead */  
   OP_ASSERTBACK,     /* 89 Positive lookbehind */  
   OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */  
   OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */  
   
   /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,  
   as there's a test for >= ONCE for a subpattern that isn't an assertion. */  
   
   OP_ONCE,           /* 92 Atomic group */  
   OP_BRA,            /* 93 Start of non-capturing bracket */  
   OP_CBRA,           /* 94 Start of capturing bracket */  
   OP_COND,           /* 95 Conditional group */  
1539    
1540    /* These three must follow the previous three, in the same order. There's a    OP_ANY,            /* 12 Match any character except newline */
1541    check for >= SBRA to distinguish the two sets. */    OP_ALLANY,         /* 13 Match any character */
1542      OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1543      OP_NOTPROP,        /* 15 \P (not Unicode property) */
1544      OP_PROP,           /* 16 \p (Unicode property) */
1545      OP_ANYNL,          /* 17 \R (any newline sequence) */
1546      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1547      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1548      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1549      OP_VSPACE,         /* 21 \v (vertical whitespace) */
1550      OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1551      OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1552      OP_EOD,            /* 24 End of data: \z */
1553    
1554      OP_CIRC,           /* 25 Start of line - not multiline */
1555      OP_CIRCM,          /* 26 Start of line - multiline */
1556      OP_DOLL,           /* 27 End of line - not multiline */
1557      OP_DOLLM,          /* 28 End of line - multiline */
1558      OP_CHAR,           /* 29 Match one character, casefully */
1559      OP_CHARI,          /* 30 Match one character, caselessly */
1560      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1561      OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1562    
1563      /* The following sets of 13 opcodes must always be kept in step because
1564      the offset from the first one is used to generate the others. */
1565    
1566      /**** Single characters, caseful, must precede the caseless ones ****/
1567    
1568      OP_STAR,           /* 33 The maximizing and minimizing versions of */
1569      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1570      OP_PLUS,           /* 35 the minimizing one second. */
1571      OP_MINPLUS,        /* 36 */
1572      OP_QUERY,          /* 37 */
1573      OP_MINQUERY,       /* 38 */
1574    
1575      OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1576      OP_MINUPTO,        /* 40 */
1577      OP_EXACT,          /* 41 Exactly n matches */
1578    
1579      OP_POSSTAR,        /* 42 Possessified star, caseful */
1580      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1581      OP_POSQUERY,       /* 44 Posesssified query, caseful */
1582      OP_POSUPTO,        /* 45 Possessified upto, caseful */
1583    
1584      /**** Single characters, caseless, must follow the caseful ones */
1585    
1586      OP_STARI,          /* 46 */
1587      OP_MINSTARI,       /* 47 */
1588      OP_PLUSI,          /* 48 */
1589      OP_MINPLUSI,       /* 49 */
1590      OP_QUERYI,         /* 50 */
1591      OP_MINQUERYI,      /* 51 */
1592    
1593      OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1594      OP_MINUPTOI,       /* 53 */
1595      OP_EXACTI,         /* 54 */
1596    
1597      OP_POSSTARI,       /* 55 Possessified star, caseless */
1598      OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1599      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1600      OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1601    
1602      /**** The negated ones must follow the non-negated ones, and match them ****/
1603      /**** Negated single character, caseful; must precede the caseless ones ****/
1604    
1605      OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1606      OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1607      OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1608      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1609      OP_NOTQUERY,       /* 63 */
1610      OP_NOTMINQUERY,    /* 64 */
1611    
1612      OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1613      OP_NOTMINUPTO,     /* 66 */
1614      OP_NOTEXACT,       /* 67 Exactly n matches */
1615    
1616      OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1617      OP_NOTPOSPLUS,     /* 69 */
1618      OP_NOTPOSQUERY,    /* 70 */
1619      OP_NOTPOSUPTO,     /* 71 */
1620    
1621      /**** Negated single character, caseless; must follow the caseful ones ****/
1622    
1623      OP_NOTSTARI,       /* 72 */
1624      OP_NOTMINSTARI,    /* 73 */
1625      OP_NOTPLUSI,       /* 74 */
1626      OP_NOTMINPLUSI,    /* 75 */
1627      OP_NOTQUERYI,      /* 76 */
1628      OP_NOTMINQUERYI,   /* 77 */
1629    
1630      OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1631      OP_NOTMINUPTOI,    /* 79 */
1632      OP_NOTEXACTI,      /* 80 Exactly n matches */
1633    
1634      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1635      OP_NOTPOSPLUSI,    /* 82 */
1636      OP_NOTPOSQUERYI,   /* 83 */
1637      OP_NOTPOSUPTOI,    /* 84 */
1638    
1639      /**** Character types ****/
1640    
1641      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1642      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1643      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1644      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1645      OP_TYPEQUERY,      /* 89 */
1646      OP_TYPEMINQUERY,   /* 90 */
1647    
1648      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1649      OP_TYPEMINUPTO,    /* 92 */
1650      OP_TYPEEXACT,      /* 93 Exactly n matches */
1651    
1652      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1653      OP_TYPEPOSPLUS,    /* 95 */
1654      OP_TYPEPOSQUERY,   /* 96 */
1655      OP_TYPEPOSUPTO,    /* 97 */
1656    
1657      /* These are used for character classes and back references; only the
1658      first six are the same as the sets above. */
1659    
1660      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1661      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1662      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1663      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1664      OP_CRQUERY,        /* 102 */
1665      OP_CRMINQUERY,     /* 103 */
1666    
1667      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1668      OP_CRMINRANGE,     /* 105 */
1669    
1670      /* End of quantifier opcodes */
1671    
1672      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1673      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1674                                  class - the difference is relevant only when a
1675                                  character > 255 is encountered. */
1676      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1677                                  class. This does both positive and negative. */
1678      OP_REF,            /* 109 Match a back reference, casefully */
1679      OP_REFI,           /* 110 Match a back reference, caselessly */
1680      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1681      OP_CALLOUT,        /* 112 Call out to external function if provided */
1682    
1683      OP_ALT,            /* 113 Start of alternation */
1684      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1685      OP_KETRMAX,        /* 115 These two must remain together and in this */
1686      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1687      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1688    
1689      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1690      asserts must remain in order. */
1691    
1692      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1693      OP_ASSERT,         /* 119 Positive lookahead */
1694      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1695      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1696      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1697    
1698      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1699      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1700      subpattern that isn't an assertion. The POS versions must immediately follow
1701      the non-POS versions in each case. */
1702    
1703      OP_ONCE,           /* 123 Atomic group, contains captures */
1704      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1705      OP_BRA,            /* 125 Start of non-capturing bracket */
1706      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1707      OP_CBRA,           /* 127 Start of capturing bracket */
1708      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1709      OP_COND,           /* 129 Conditional group */
1710    
1711    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    /* These five must follow the previous five, in the same order. There's a
1712    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    check for >= SBRA to distinguish the two sets. */
   OP_SCOND,          /* 98 Conditional group, check empty */  
   
   OP_CREF,           /* 99 Used to hold a capture number as condition */  
   OP_RREF,           /* 100 Used to hold a recursion number as condition */  
   OP_DEF,            /* 101 The DEFINE condition */  
1713    
1714    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1715    OP_BRAMINZERO,     /* 103 order. */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1716      OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1717      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1718      OP_SCOND,          /* 134 Conditional group, check empty */
1719    
1720      /* The next two pairs must (respectively) be kept together. */
1721    
1722      OP_CREF,           /* 135 Used to hold a capture number as condition */
1723      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1724      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1725      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1726      OP_DEF,            /* 139 The DEFINE condition */
1727    
1728      OP_BRAZERO,        /* 140 These two must remain together and in this */
1729      OP_BRAMINZERO,     /* 141 order. */
1730      OP_BRAPOSZERO,     /* 142 */
1731    
1732    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1733    
1734    OP_PRUNE,          /* 104 */    OP_MARK,           /* 143 always has an argument */
1735    OP_SKIP,           /* 105 */    OP_PRUNE,          /* 144 */
1736    OP_THEN,           /* 106 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1737    OP_COMMIT,         /* 107 */    OP_SKIP,           /* 146 */
1738      OP_SKIP_ARG,       /* 147 same, but with argument */
1739      OP_THEN,           /* 148 */
1740      OP_THEN_ARG,       /* 149 same, but with argument */
1741      OP_COMMIT,         /* 150 */
1742    
1743    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1744    
1745    OP_FAIL,           /* 108 */    OP_FAIL,           /* 151 */
1746    OP_ACCEPT,         /* 109 */    OP_ACCEPT,         /* 152 */
1747      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1748      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1749    
1750    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1751    
1752    OP_SKIPZERO        /* 110 */    OP_SKIPZERO,       /* 155 */
1753    
1754      /* This is not an opcode, but is used to check that tables indexed by opcode
1755      are the correct length, in order to catch updating errors - there have been
1756      some in the past. */
1757    
1758      OP_TABLE_LENGTH
1759  };  };
1760    
1761    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1762    definitions that follow must also be updated to match. There are also tables
1763    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1764    
1765    
1766  /* 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
1767  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
1768    only in pcre_printint.c, which fills out the full names in many cases (and in
1769    some cases doesn't actually use these names at all). */
1770    
1771  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1772    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1773    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1774    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1775    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1776    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1777    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1778      "{", "{", "{",                                                  \
1779    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1780    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1781      "{", "{", "{",                                                  \
1782      "*+","++", "?+", "{",                                           \
1783      "*", "*?", "+", "+?", "?", "??",                                \
1784      "{", "{", "{",                                                  \
1785      "*+","++", "?+", "{",                                           \
1786      "*", "*?", "+", "+?", "?", "??",                                \
1787      "{", "{", "{",                                                  \
1788    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1789    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1790    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1791    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1792    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1793    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1794    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1795    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1796    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once", "Once_NC",                                              \
1797    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Bra", "BraPos", "CBra", "CBraPos",                             \
1798    "Skip zero"    "Cond",                                                         \
1799      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1800      "SCond",                                                        \
1801      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1802      "Brazero", "Braminzero", "Braposzero",                          \
1803      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1804      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1805      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1806      "Close", "Skip zero"
1807    
1808    
1809  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 819  in UTF-8 mode. The code that uses this t Line 1819  in UTF-8 mode. The code that uses this t
1819    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1820    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1821    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1822    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1823    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1824    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1825    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1826      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1827    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1828    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1829    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1830    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1831    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1832    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1833    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1834      2+IMM2_SIZE,                   /* exact                                  */ \
1835      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1836      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1837      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1838      2+IMM2_SIZE,                   /* exact I                                */ \
1839      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1840    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1841    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1842    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1843    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1844      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1845      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1846      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1847      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1848      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1849    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1850    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1851    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1852    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
1853      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1854    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1855    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1856    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1857   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
1858   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
1859    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1860    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1861      1+IMM2_SIZE,                   /* REFI                                   */ \
1862    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1863    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1864    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1865    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1866    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1867    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1868      1+LINK_SIZE,                   /* KetRpos                                */ \
1869      1+LINK_SIZE,                   /* Reverse                                */ \
1870    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1871    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1872    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1873    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1874    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1875      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1876    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1877    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
1878      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1879      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1880    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1881    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1882    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1883      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1884      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1885    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1886    3,                             /* CREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1887    3,                             /* RREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1888    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1889    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1890    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1891    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1892      1, 3,                          /* THEN, THEN_ARG                         */ \
1893      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1894      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1895    
1896    /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1897  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  condition. */
1898    
1899  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1900    
1901  /* 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
1902  tracked. */  easily be tracked. When a new number is added, the table called eint in
1903    pcreposix.c must be updated. */
1904    
1905  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1906         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 884  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1908  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1908         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1909         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1910         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1911         ERR60, ERR61, ERR62, ERR63 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1912           ERR70, ERR71, ERRCOUNT };
1913    
1914  /* 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
1915  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 900  Because people can now save and re-use c Line 1925  Because people can now save and re-use c
1925  structure should be made at the end, and something earlier (e.g. a new  structure should be made at the end, and something earlier (e.g. a new
1926  flag in the options or one of the dummy fields) should indicate that the new  flag in the options or one of the dummy fields) should indicate that the new
1927  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1928  NOTE NOTE NOTE:  NOTE NOTE NOTE
1929  */  */
1930    
1931  typedef struct real_pcre {  typedef struct real_pcre {
# Line 911  typedef struct real_pcre { Line 1936  typedef struct real_pcre {
1936    pcre_uint16 dummy1;             /* For future use */    pcre_uint16 dummy1;             /* For future use */
1937    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1938    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1939    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
1940    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
1941    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
1942    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
1943    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1944    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1945    
1946    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1947    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1948  } real_pcre;  } real_pcre;
1949    
1950  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
# Line 927  remark (see NOTE above) about extending Line 1952  remark (see NOTE above) about extending
1952    
1953  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1954    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1955    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1956    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
1957      pcre_uint32 minlength;          /* Minimum subject length */
1958  } pcre_study_data;  } pcre_study_data;
1959    
1960    /* Structure for building a chain of open capturing subpatterns during
1961    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1962    encountered. This is also used to identify subpatterns that contain recursive
1963    back references to themselves, so that they can be made atomic. */
1964    
1965    typedef struct open_capitem {
1966      struct open_capitem *next;    /* Chain link */
1967      pcre_uint16 number;           /* Capture number */
1968      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1969    } open_capitem;
1970    
1971  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1972  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1973    
1974  typedef struct compile_data {  typedef struct compile_data {
1975    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1976    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1977    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1978    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1979    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1980    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1981    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1982    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1983    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
1984    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
1985    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
1986    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
1987    int  bracount;                /* Count of capturing parens as we compile */    int  name_entry_size;             /* Size of each entry */
1988    int  final_bracount;          /* Saved value after first pass */    int  bracount;                    /* Count of capturing parens as we compile */
1989    int  top_backref;             /* Maximum back reference */    int  final_bracount;              /* Saved value after first pass */
1990    unsigned int backref_map;     /* Bitmap of low back refs */    int  top_backref;                 /* Maximum back reference */
1991    int  external_options;        /* External (initial) options */    unsigned int backref_map;         /* Bitmap of low back refs */
1992    int  external_flags;          /* External flag bits to be set */    int  assert_depth;                /* Depth of nested assertions */
1993    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_options;            /* External (initial) options */
1994    BOOL had_accept;              /* (*ACCEPT) encountered */    int  external_flags;              /* External flag bits to be set */
1995    int  nltype;                  /* Newline type */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1996    int  nllen;                   /* Newline string length */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1997    uschar nl[4];                 /* Newline string when fixed length */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1998      int  nltype;                      /* Newline type */
1999      int  nllen;                       /* Newline string length */
2000      pcre_uchar nl[4];                 /* Newline string when fixed length */
2001  } compile_data;  } compile_data;
2002    
2003  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
2004  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
2005    
2006  typedef struct branch_chain {  typedef struct branch_chain {
2007    struct branch_chain *outer;    struct branch_chain *outer;
2008    uschar *current;    pcre_uchar *current_branch;
2009  } branch_chain;  } branch_chain;
2010    
2011  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2012  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2013    
2014  typedef struct recursion_info {  typedef struct recursion_info {
2015    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2016    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2017    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2018    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
2019    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 */  
2020  } recursion_info;  } recursion_info;
2021    
2022    /* A similar structure for pcre_dfa_exec(). */
2023    
2024    typedef struct dfa_recursion_info {
2025      struct dfa_recursion_info *prevrec;
2026      int group_num;
2027      PCRE_PUCHAR subject_position;
2028    } dfa_recursion_info;
2029    
2030  /* 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
2031  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
2032  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
2033    pcre_exec(). */
2034    
2035  typedef struct eptrblock {  typedef struct eptrblock {
2036    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2037    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2038  } eptrblock;  } eptrblock;
2039    
2040    
# Line 997  typedef struct match_data { Line 2045  typedef struct match_data {
2045    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2046    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2047    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2048    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2049    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2050    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2051    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2052    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2053    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
2054    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
2055    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
2056    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
2057    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2058    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2059    BOOL   utf8;                  /* UTF8 flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2060    BOOL   endonly;               /* Dollar not before final \n */    BOOL   offset_overflow;         /* Set if too many extractions */
2061    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notbol;                  /* NOTBOL flag */
2062    BOOL   partial;               /* PARTIAL flag */    BOOL   noteol;                  /* NOTEOL flag */
2063    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2064    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2065    const uschar *start_code;     /* For use when recursing */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2066    USPTR  start_subject;         /* Start of the subject string */    BOOL   endonly;                 /* Dollar not before final \n */
2067    USPTR  end_subject;           /* End of the subject string */    BOOL   notempty;                /* Empty string match not wanted */
2068    USPTR  start_match_ptr;       /* Start of matched string */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2069    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2070    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2071    int    capture_last;          /* Most recent capture number */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2072    int    start_offset;          /* The start offset value */    const  pcre_uchar *start_code;  /* For use when recursing */
2073    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2074    int    eptrn;                 /* Next free eptrblock */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2075    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2076    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2077      PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2078      int    partial;                 /* PARTIAL options */
2079      int    end_offset_top;          /* Highwater mark at end of match */
2080      int    capture_last;            /* Most recent capture number */
2081      int    start_offset;            /* The start offset value */
2082      int    match_function_type;     /* Set for certain special calls of MATCH() */
2083      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2084      int    eptrn;                   /* Next free eptrblock */
2085      recursion_info *recursive;      /* Linked list of recursion data */
2086      void  *callout_data;            /* To pass back to callouts */
2087      const  pcre_uchar *mark;        /* Mark pointer to pass back */
2088      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2089  } match_data;  } match_data;
2090    
2091  /* 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
2092  functions. */  functions. */
2093    
2094  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2095    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2096    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2097    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2098    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2099    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
2100    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
2101    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
2102    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
2103    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
2104    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
2105      pcre_uchar nl[4];                 /* Newline string when fixed */
2106      void *callout_data;               /* To pass back to callouts */
2107      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2108  } dfa_match_data;  } dfa_match_data;
2109    
2110  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1077  total length. */ Line 2140  total length. */
2140  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2141  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2142    
2143    /* Internal function prefix */
2144    
2145    #ifdef COMPILE_PCRE8
2146    #define PRIV(name) _pcre_##name
2147    #else
2148    #ifdef COMPILE_PCRE16
2149    #define PRIV(name) _pcre16_##name
2150    #else
2151    #error Unsupported compiling mode
2152    #endif /* COMPILE_PCRE16 */
2153    #endif /* COMPILE_PCRE8 */
2154    
2155  /* 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
2156  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
2157  relocations in shared libraries, it now has an offset into a single string  relocations in shared libraries, it now has an offset into a single string
# Line 1094  of the exported public functions. They h Line 2169  of the exported public functions. They h
2169  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
2170  pcre_tables.c module. */  pcre_tables.c module. */
2171    
2172  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
 extern const int    _pcre_utf8_table2[];  
 extern const int    _pcre_utf8_table3[];  
 extern const uschar _pcre_utf8_table4[];  
2173    
2174  extern const int    _pcre_utf8_table1_size;  extern const int            PRIV(utf8_table1)[];
2175    extern const int            PRIV(utf8_table1_size);
2176    extern const int            PRIV(utf8_table2)[];
2177    extern const int            PRIV(utf8_table3)[];
2178    extern const pcre_uint8     PRIV(utf8_table4)[];
2179    
2180  extern const char   _pcre_utt_names[];  #endif /* COMPILE_PCRE8 */
 extern const ucp_type_table _pcre_utt[];  
 extern const int _pcre_utt_size;  
2181    
2182  extern const uschar _pcre_default_tables[];  extern const char           PRIV(utt_names)[];
2183    extern const ucp_type_table PRIV(utt)[];
2184    extern const int            PRIV(utt_size);
2185    
2186  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(default_tables)[];
2187    
2188    extern const pcre_uint8     PRIV(OP_lengths)[];
2189    
2190    
2191  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2192  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
2193  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2194    
2195  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  /* String comparison functions. */
2196                        int *, BOOL);  #ifdef COMPILE_PCRE8
2197  extern int          _pcre_ord2utf8(int, uschar *);  
2198  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2199                        const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2200  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
2201  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
2202  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
2203  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,    strncmp((char *)(str1), (char *)(str2), (num))
2204                        int *, BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
2205  extern BOOL         _pcre_xclass(int, const uschar *);    strncmp((char *)(str1), (str2), (num))
2206    #define STRLEN_UC(str) strlen((const char *)str)
2207    
2208    #else
2209    
2210    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2211                               const pcre_uchar *);
2212    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2213                               const char *);
2214    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2215                               const pcre_uchar *, unsigned int num);
2216    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2217                               const char *, unsigned int num);
2218    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2219    
2220    #define STRCMP_UC_UC(str1, str2) \
2221      PRIV(strcmp_uc_uc)((str1), (str2))
2222    #define STRCMP_UC_C8(str1, str2) \
2223      PRIV(strcmp_uc_c8)((str1), (str2))
2224    #define STRNCMP_UC_UC(str1, str2, num) \
2225      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2226    #define STRNCMP_UC_C8(str1, str2, num) \
2227      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2228    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2229    
2230    #endif /* COMPILE_PCRE8 */
2231    
2232    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2233    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2234                               int *, BOOL);
2235    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2236    extern real_pcre        *PRIV(try_flipped)(const real_pcre *, real_pcre *,
2237                               const pcre_study_data *, pcre_study_data *);
2238    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2239    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2240                               int *, BOOL);
2241    extern BOOL              PRIV(xclass)(int, const pcre_uchar *);
2242    
2243    #ifdef SUPPORT_JIT
2244    extern void              PRIV(jit_compile)(const real_pcre *, pcre_extra *);
2245    extern int               PRIV(jit_exec)(const real_pcre *, void *,
2246                               const pcre_uchar *, int, int, int, int, int *, int);
2247    extern void              PRIV(jit_free)(void *);
2248    #endif
2249    
2250    /* Unicode character database (UCD) */
2251    
2252    typedef struct {
2253      pcre_uint8 script;
2254      pcre_uint8 chartype;
2255      pcre_int32 other_case;
2256    } ucd_record;
2257    
2258    extern const ucd_record  PRIV(ucd_records)[];
2259    extern const pcre_uint8  PRIV(ucd_stage1)[];
2260    extern const pcre_uint16 PRIV(ucd_stage2)[];
2261    extern const int         PRIV(ucp_gentype)[];
2262    #ifdef SUPPORT_JIT
2263    extern const int         PRIV(ucp_typerange)[];
2264    #endif
2265    
2266    #ifdef SUPPORT_UCP
2267    /* UCD access macros */
2268    
2269    #define UCD_BLOCK_SIZE 128
2270    #define GET_UCD(ch) (PRIV(ucd_records) + \
2271            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2272            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2273    
2274    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2275    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2276    #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2277    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2278    
2279    #endif /* SUPPORT_UCP */
2280    
2281  #endif  #endif
2282    

Legend:
Removed from v.335  
changed lines
  Added in v.789

  ViewVC Help
Powered by ViewVC 1.1.5