/[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 763 by zherczeg, Tue Nov 22 21:46:22 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 45  functions whose names all begin with "_p Line 45  functions whose names all begin with "_p
45  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
46  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
49    
50  #if 0  #if 0
51  #define DEBUG  #define PCRE_DEBUG
52    #endif
53    
54    /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
55    #ifndef COMPILE_PCRE16
56    #define COMPILE_PCRE8
57    #endif
58    
59    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
60    script prevents both being selected, but not everybody uses "configure". */
61    
62    #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
63    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
64    #endif
65    
66    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
67    "configure" script ensures this, but not everybody uses "configure". */
68    
69    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
70    #define SUPPORT_UTF8 1
71  #endif  #endif
72    
73  /* 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 79  It turns out that the Mac Debugging.h he
79  be absolutely sure we get our version. */  be absolutely sure we get our version. */
80    
81  #undef DPRINTF  #undef DPRINTF
82  #ifdef DEBUG  #ifdef PCRE_DEBUG
83  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
84  #else  #else
85  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
# Line 72  setjmp and stdarg are used is when NO_RE Line 91  setjmp and stdarg are used is when NO_RE
91    
92  #include <ctype.h>  #include <ctype.h>
93  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
94  #include <stddef.h>  #include <stddef.h>
95  #include <stdio.h>  #include <stdio.h>
96  #include <stdlib.h>  #include <stdlib.h>
# Line 132  PCRE_EXP_DATA_DEFN only if they are not Line 149  PCRE_EXP_DATA_DEFN only if they are not
149  #  endif  #  endif
150  #endif  #endif
151    
152  /* 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
153    a "calling convention" before exported function names. (This is secondhand
154    information; I know nothing about MSVC myself). For example, something like
155    
156      void __cdecl function(....)
157    
158    might be needed. In order so make this easy, all the exported functions have
159    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
160    set, we ensure here that it has no effect. */
161    
162    #ifndef PCRE_CALL_CONVENTION
163    #define PCRE_CALL_CONVENTION
164    #endif
165    
166    /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
167  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
168  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
169  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
170  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
171    
172    typedef unsigned char pcre_uint8;
173    
174  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
175    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
176      typedef short pcre_int16;
177  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
178    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
179      typedef int pcre_int16;
180  #else  #else
181    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
182  #endif  #endif
183    
184  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
185    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
186      typedef int pcre_int32;
187  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
188    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
189      typedef long int pcre_int32;
190  #else  #else
191    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
192  #endif  #endif
193    
194    /* When checking for integer overflow in pcre_compile(), we need to handle
195    large integers. If a 64-bit integer type is available, we can use that.
196    Otherwise we have to cast to double, which of course requires floating point
197    arithmetic. Handle this by defining a macro for the appropriate type. If
198    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
199    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
200    by "configure". */
201    
202    #if HAVE_STDINT_H
203    #include <stdint.h>
204    #elif HAVE_INTTYPES_H
205    #include <inttypes.h>
206    #endif
207    
208    #if defined INT64_MAX || defined int64_t
209    #define INT64_OR_DOUBLE int64_t
210    #else
211    #define INT64_OR_DOUBLE double
212    #endif
213    
214  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
215  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
216  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 *,
217  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. We define a short type
218  to save lots of typing. I tried "uchar", but it causes problems on Digital  for the current character representation (either 8 or 16 bit) to save lots
219  Unix, where it is defined in sys/types, so use "uschar" instead. */  of typing. I tried "uchar", but it causes problems on Digital Unix, where
220    it is defined in sys/types, so use "uschar" instead. */
221    
222  typedef unsigned char uschar;  #ifndef COMPILE_PCRE16
223    typedef unsigned char pcre_uchar;
224    #else
225    #if USHRT_MAX != 65535
226    /* This is a warning message. Change PCRE_SCHAR16 to a 16 bit data type in
227    pcre.h(.in) and disable (comment out) this message. */
228    #error Warning: PCRE_SCHAR16 is not a 16 bit data type.
229    #endif
230    typedef pcre_uint16 pcre_uchar;
231    #endif
232    
233  /* 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
234  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 217  used for the external interface and appe Line 284  used for the external interface and appe
284  must begin with PCRE_. */  must begin with PCRE_. */
285    
286  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
287  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
288  #else  #else
289  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
 #endif  
   
290    
291    /* PCRE_SPTR is defined in pcre.h. */
292    #define USPTR const uschar *
293    #endif
294    
295  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
296  values. */  values. */
# Line 346  is automated on Unix systems via the "co Line 412  is automated on Unix systems via the "co
412  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
413  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
414    
415    #ifdef COMPILE_PCRE8
416    
417    #define IMM2_SIZE 2
418    
419  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
420    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
421    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 353  capturing parenthesis numbers in back re Line 423  capturing parenthesis numbers in back re
423  #define GET2(a,n) \  #define GET2(a,n) \
424    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
425    
 #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  
426    
427    #elif defined(COMPILE_PCRE16)
428    
429    #define IMM2_SIZE 1
430    
431    #define PUT2(a,n,d)   \
432       a[n] = d
433    
434    #define GET2(a,n) \
435       a[n]
436    
437    #else
438    #error Unsupported compiling mode
439    #endif /* COMPILE_PCRE8 */
440    
441    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
442    
443  /* 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
444  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
445  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
446  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
447  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
448    UTF-8 support is omitted, we don't even define them. */
449    
450  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
 #define NEXTCHAR(p) p++;  
451  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
452  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
453  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
454  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
455  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
456    /* #define GETCHARLENTEST(c, eptr, len) */
457  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
458    
459  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
460    
461  /* Advance a character pointer one byte in non-UTF-8 mode and by one character  /* These macros were originally written in the form of loops that used data
462  in UTF-8 mode. */  from the tables whose names start with _pcre_utf8_table. They were rewritten by
463    a user so as not to use loops, because in some environments this gives a
464    significant performance advantage, and it seems never to do any harm. */
465    
466  #define NEXTCHAR(p) \  /* Base macro to pick up the remaining bytes of a UTF-8 character, not
467    p++; \  advancing the pointer. */
468    if (utf8) { while((*p & 0xc0) == 0x80) p++; }  
469    #define GETUTF8(c, eptr) \
470        { \
471        if ((c & 0x20) == 0) \
472          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
473        else if ((c & 0x10) == 0) \
474          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
475        else if ((c & 0x08) == 0) \
476          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
477          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
478        else if ((c & 0x04) == 0) \
479          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
480              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
481              (eptr[4] & 0x3f); \
482        else \
483          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
484              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
485              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
486        }
487    
488  /* 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
489  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
490    
491  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
492    c = *eptr; \    c = *eptr; \
493    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; \  
       } \  
     }  
494    
495  /* 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
496  pointer. */  pointer. */
497    
498  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
499    c = *eptr; \    c = *eptr; \
500    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
501    
502    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
503    the pointer. */
504    
505    #define GETUTF8INC(c, eptr) \
506      { \      { \
507      int gcii; \      if ((c & 0x20) == 0) \
508      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
509      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
510      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
511      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
512          eptr += 2; \
513          } \
514        else if ((c & 0x08) == 0) \
515        { \        { \
516        gcss -= 6; \        c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
517        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
518          eptr += 3; \
519          } \
520        else if ((c & 0x04) == 0) \
521          { \
522          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
523              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
524              (eptr[3] & 0x3f); \
525          eptr += 4; \
526          } \
527        else \
528          { \
529          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
530              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
531              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
532          eptr += 5; \
533        } \        } \
534      }      }
535    
# Line 421  know we are in UTF-8 mode. */ Line 538  know we are in UTF-8 mode. */
538    
539  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
540    c = *eptr++; \    c = *eptr++; \
541    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; \  
       } \  
     }  
542    
543  /* 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.
544    This is called when we don't know if we are in UTF-8 mode. */
545    
546  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
547    c = *eptr++; \    c = *eptr++; \
548    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
549    
550    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
551    advancing the pointer, incrementing the length. */
552    
553    #define GETUTF8LEN(c, eptr, len) \
554      { \      { \
555      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
556      int gcss = 6*gcaa; \        { \
557      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
558      while (gcaa-- > 0) \        len++; \
559          } \
560        else if ((c & 0x10)  == 0) \
561          { \
562          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
563          len += 2; \
564          } \
565        else if ((c & 0x08)  == 0) \
566          {\
567          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
568              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
569          len += 3; \
570          } \
571        else if ((c & 0x04)  == 0) \
572        { \        { \
573        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
574        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
575              (eptr[4] & 0x3f); \
576          len += 4; \
577          } \
578        else \
579          {\
580          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
581              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
582              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
583          len += 5; \
584        } \        } \
585      }      }
586    
# Line 454  if there are extra bytes. This is called Line 589  if there are extra bytes. This is called
589    
590  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
591    c = *eptr; \    c = *eptr; \
592    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
593      { \  
594      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
595      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
596      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
597      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
598      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
599        { \    c = *eptr; \
600        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
601    
602  /* 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
603  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 605  because almost all calls are already wit
605    
606  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
607    
608  #endif  #endif  /* SUPPORT_UTF8 */
609    
610    
611  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
# Line 485  Standard C system should have one. */ Line 616  Standard C system should have one. */
616  #endif  #endif
617    
618    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
619  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
620  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
621  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
622    the restrictions on partial matching have been lifted. It remains for backwards
623    compatibility. */
624    
625  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
626  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 499  are in a 16-bit flags word. */ Line 628  are in a 16-bit flags word. */
628  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
629  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
630  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
631    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
632    
633  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
634    
635  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
636    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
637    
638  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
639  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 510  time, run time, or study time, respectiv Line 641  time, run time, or study time, respectiv
641  #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| \
642                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
643    
644  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
645    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
646     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
647     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| \
648     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
649       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
650    
651  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
652    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
653     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
654       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
655    
656  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
657    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
658     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
659     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
660       PCRE_NO_START_OPTIMIZE)
661    
662  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
663       PCRE_STUDY_JIT_COMPILE
664    
665  /* 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
666  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 548  variable-length repeat, or a anything ot Line 683  variable-length repeat, or a anything ot
683  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
684  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
685    
686  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
687    environments where these macros are defined elsewhere. Unfortunately, there
688    is no way to do the same for the typedef. */
689    
690  typedef int BOOL;  typedef int BOOL;
691    
692    #ifndef FALSE
693  #define FALSE   0  #define FALSE   0
694  #define TRUE    1  #define TRUE    1
695    #endif
696    
697    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
698    character constants like '*' because the compiler would emit their EBCDIC code,
699    which is different from their ASCII/UTF-8 code. Instead we define macros for
700    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
701    is enabled. When UTF-8 support is not enabled, the definitions use character
702    literals. Both character and string versions of each character are needed, and
703    there are some longer strings as well.
704    
705    This means that, on EBCDIC platforms, the PCRE library can handle either
706    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
707    would need different lookups depending on whether PCRE_UTF8 was set or not.
708    This would make it impossible to use characters in switch/case statements,
709    which would reduce performance. For a theoretical use (which nobody has asked
710    for) in a minority area (EBCDIC platforms), this is not sensible. Any
711    application that did need both could compile two versions of the library, using
712    macros to give the functions distinct names. */
713    
714    #ifndef SUPPORT_UTF8
715    
716    /* UTF-8 support is not enabled; use the platform-dependent character literals
717    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
718    
719    #define CHAR_HT                     '\t'
720    #define CHAR_VT                     '\v'
721    #define CHAR_FF                     '\f'
722    #define CHAR_CR                     '\r'
723    #define CHAR_NL                     '\n'
724    #define CHAR_BS                     '\b'
725    #define CHAR_BEL                    '\a'
726    #ifdef EBCDIC
727    #define CHAR_ESC                    '\047'
728    #define CHAR_DEL                    '\007'
729    #else
730    #define CHAR_ESC                    '\033'
731    #define CHAR_DEL                    '\177'
732    #endif
733    
734    #define CHAR_SPACE                  ' '
735    #define CHAR_EXCLAMATION_MARK       '!'
736    #define CHAR_QUOTATION_MARK         '"'
737    #define CHAR_NUMBER_SIGN            '#'
738    #define CHAR_DOLLAR_SIGN            '$'
739    #define CHAR_PERCENT_SIGN           '%'
740    #define CHAR_AMPERSAND              '&'
741    #define CHAR_APOSTROPHE             '\''
742    #define CHAR_LEFT_PARENTHESIS       '('
743    #define CHAR_RIGHT_PARENTHESIS      ')'
744    #define CHAR_ASTERISK               '*'
745    #define CHAR_PLUS                   '+'
746    #define CHAR_COMMA                  ','
747    #define CHAR_MINUS                  '-'
748    #define CHAR_DOT                    '.'
749    #define CHAR_SLASH                  '/'
750    #define CHAR_0                      '0'
751    #define CHAR_1                      '1'
752    #define CHAR_2                      '2'
753    #define CHAR_3                      '3'
754    #define CHAR_4                      '4'
755    #define CHAR_5                      '5'
756    #define CHAR_6                      '6'
757    #define CHAR_7                      '7'
758    #define CHAR_8                      '8'
759    #define CHAR_9                      '9'
760    #define CHAR_COLON                  ':'
761    #define CHAR_SEMICOLON              ';'
762    #define CHAR_LESS_THAN_SIGN         '<'
763    #define CHAR_EQUALS_SIGN            '='
764    #define CHAR_GREATER_THAN_SIGN      '>'
765    #define CHAR_QUESTION_MARK          '?'
766    #define CHAR_COMMERCIAL_AT          '@'
767    #define CHAR_A                      'A'
768    #define CHAR_B                      'B'
769    #define CHAR_C                      'C'
770    #define CHAR_D                      'D'
771    #define CHAR_E                      'E'
772    #define CHAR_F                      'F'
773    #define CHAR_G                      'G'
774    #define CHAR_H                      'H'
775    #define CHAR_I                      'I'
776    #define CHAR_J                      'J'
777    #define CHAR_K                      'K'
778    #define CHAR_L                      'L'
779    #define CHAR_M                      'M'
780    #define CHAR_N                      'N'
781    #define CHAR_O                      'O'
782    #define CHAR_P                      'P'
783    #define CHAR_Q                      'Q'
784    #define CHAR_R                      'R'
785    #define CHAR_S                      'S'
786    #define CHAR_T                      'T'
787    #define CHAR_U                      'U'
788    #define CHAR_V                      'V'
789    #define CHAR_W                      'W'
790    #define CHAR_X                      'X'
791    #define CHAR_Y                      'Y'
792    #define CHAR_Z                      'Z'
793    #define CHAR_LEFT_SQUARE_BRACKET    '['
794    #define CHAR_BACKSLASH              '\\'
795    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
796    #define CHAR_CIRCUMFLEX_ACCENT      '^'
797    #define CHAR_UNDERSCORE             '_'
798    #define CHAR_GRAVE_ACCENT           '`'
799    #define CHAR_a                      'a'
800    #define CHAR_b                      'b'
801    #define CHAR_c                      'c'
802    #define CHAR_d                      'd'
803    #define CHAR_e                      'e'
804    #define CHAR_f                      'f'
805    #define CHAR_g                      'g'
806    #define CHAR_h                      'h'
807    #define CHAR_i                      'i'
808    #define CHAR_j                      'j'
809    #define CHAR_k                      'k'
810    #define CHAR_l                      'l'
811    #define CHAR_m                      'm'
812    #define CHAR_n                      'n'
813    #define CHAR_o                      'o'
814    #define CHAR_p                      'p'
815    #define CHAR_q                      'q'
816    #define CHAR_r                      'r'
817    #define CHAR_s                      's'
818    #define CHAR_t                      't'
819    #define CHAR_u                      'u'
820    #define CHAR_v                      'v'
821    #define CHAR_w                      'w'
822    #define CHAR_x                      'x'
823    #define CHAR_y                      'y'
824    #define CHAR_z                      'z'
825    #define CHAR_LEFT_CURLY_BRACKET     '{'
826    #define CHAR_VERTICAL_LINE          '|'
827    #define CHAR_RIGHT_CURLY_BRACKET    '}'
828    #define CHAR_TILDE                  '~'
829    
830    #define STR_HT                      "\t"
831    #define STR_VT                      "\v"
832    #define STR_FF                      "\f"
833    #define STR_CR                      "\r"
834    #define STR_NL                      "\n"
835    #define STR_BS                      "\b"
836    #define STR_BEL                     "\a"
837    #ifdef EBCDIC
838    #define STR_ESC                     "\047"
839    #define STR_DEL                     "\007"
840    #else
841    #define STR_ESC                     "\033"
842    #define STR_DEL                     "\177"
843    #endif
844    
845    #define STR_SPACE                   " "
846    #define STR_EXCLAMATION_MARK        "!"
847    #define STR_QUOTATION_MARK          "\""
848    #define STR_NUMBER_SIGN             "#"
849    #define STR_DOLLAR_SIGN             "$"
850    #define STR_PERCENT_SIGN            "%"
851    #define STR_AMPERSAND               "&"
852    #define STR_APOSTROPHE              "'"
853    #define STR_LEFT_PARENTHESIS        "("
854    #define STR_RIGHT_PARENTHESIS       ")"
855    #define STR_ASTERISK                "*"
856    #define STR_PLUS                    "+"
857    #define STR_COMMA                   ","
858    #define STR_MINUS                   "-"
859    #define STR_DOT                     "."
860    #define STR_SLASH                   "/"
861    #define STR_0                       "0"
862    #define STR_1                       "1"
863    #define STR_2                       "2"
864    #define STR_3                       "3"
865    #define STR_4                       "4"
866    #define STR_5                       "5"
867    #define STR_6                       "6"
868    #define STR_7                       "7"
869    #define STR_8                       "8"
870    #define STR_9                       "9"
871    #define STR_COLON                   ":"
872    #define STR_SEMICOLON               ";"
873    #define STR_LESS_THAN_SIGN          "<"
874    #define STR_EQUALS_SIGN             "="
875    #define STR_GREATER_THAN_SIGN       ">"
876    #define STR_QUESTION_MARK           "?"
877    #define STR_COMMERCIAL_AT           "@"
878    #define STR_A                       "A"
879    #define STR_B                       "B"
880    #define STR_C                       "C"
881    #define STR_D                       "D"
882    #define STR_E                       "E"
883    #define STR_F                       "F"
884    #define STR_G                       "G"
885    #define STR_H                       "H"
886    #define STR_I                       "I"
887    #define STR_J                       "J"
888    #define STR_K                       "K"
889    #define STR_L                       "L"
890    #define STR_M                       "M"
891    #define STR_N                       "N"
892    #define STR_O                       "O"
893    #define STR_P                       "P"
894    #define STR_Q                       "Q"
895    #define STR_R                       "R"
896    #define STR_S                       "S"
897    #define STR_T                       "T"
898    #define STR_U                       "U"
899    #define STR_V                       "V"
900    #define STR_W                       "W"
901    #define STR_X                       "X"
902    #define STR_Y                       "Y"
903    #define STR_Z                       "Z"
904    #define STR_LEFT_SQUARE_BRACKET     "["
905    #define STR_BACKSLASH               "\\"
906    #define STR_RIGHT_SQUARE_BRACKET    "]"
907    #define STR_CIRCUMFLEX_ACCENT       "^"
908    #define STR_UNDERSCORE              "_"
909    #define STR_GRAVE_ACCENT            "`"
910    #define STR_a                       "a"
911    #define STR_b                       "b"
912    #define STR_c                       "c"
913    #define STR_d                       "d"
914    #define STR_e                       "e"
915    #define STR_f                       "f"
916    #define STR_g                       "g"
917    #define STR_h                       "h"
918    #define STR_i                       "i"
919    #define STR_j                       "j"
920    #define STR_k                       "k"
921    #define STR_l                       "l"
922    #define STR_m                       "m"
923    #define STR_n                       "n"
924    #define STR_o                       "o"
925    #define STR_p                       "p"
926    #define STR_q                       "q"
927    #define STR_r                       "r"
928    #define STR_s                       "s"
929    #define STR_t                       "t"
930    #define STR_u                       "u"
931    #define STR_v                       "v"
932    #define STR_w                       "w"
933    #define STR_x                       "x"
934    #define STR_y                       "y"
935    #define STR_z                       "z"
936    #define STR_LEFT_CURLY_BRACKET      "{"
937    #define STR_VERTICAL_LINE           "|"
938    #define STR_RIGHT_CURLY_BRACKET     "}"
939    #define STR_TILDE                   "~"
940    
941    #define STRING_ACCEPT0              "ACCEPT\0"
942    #define STRING_COMMIT0              "COMMIT\0"
943    #define STRING_F0                   "F\0"
944    #define STRING_FAIL0                "FAIL\0"
945    #define STRING_MARK0                "MARK\0"
946    #define STRING_PRUNE0               "PRUNE\0"
947    #define STRING_SKIP0                "SKIP\0"
948    #define STRING_THEN                 "THEN"
949    
950    #define STRING_alpha0               "alpha\0"
951    #define STRING_lower0               "lower\0"
952    #define STRING_upper0               "upper\0"
953    #define STRING_alnum0               "alnum\0"
954    #define STRING_ascii0               "ascii\0"
955    #define STRING_blank0               "blank\0"
956    #define STRING_cntrl0               "cntrl\0"
957    #define STRING_digit0               "digit\0"
958    #define STRING_graph0               "graph\0"
959    #define STRING_print0               "print\0"
960    #define STRING_punct0               "punct\0"
961    #define STRING_space0               "space\0"
962    #define STRING_word0                "word\0"
963    #define STRING_xdigit               "xdigit"
964    
965    #define STRING_DEFINE               "DEFINE"
966    
967    #define STRING_CR_RIGHTPAR             "CR)"
968    #define STRING_LF_RIGHTPAR             "LF)"
969    #define STRING_CRLF_RIGHTPAR           "CRLF)"
970    #define STRING_ANY_RIGHTPAR            "ANY)"
971    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
972    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
973    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
974    #define STRING_UTF8_RIGHTPAR           "UTF8)"
975    #define STRING_UCP_RIGHTPAR            "UCP)"
976    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
977    
978    #else  /* SUPPORT_UTF8 */
979    
980    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
981    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
982    only. */
983    
984    #define CHAR_HT                     '\011'
985    #define CHAR_VT                     '\013'
986    #define CHAR_FF                     '\014'
987    #define CHAR_CR                     '\015'
988    #define CHAR_NL                     '\012'
989    #define CHAR_BS                     '\010'
990    #define CHAR_BEL                    '\007'
991    #define CHAR_ESC                    '\033'
992    #define CHAR_DEL                    '\177'
993    
994    #define CHAR_SPACE                  '\040'
995    #define CHAR_EXCLAMATION_MARK       '\041'
996    #define CHAR_QUOTATION_MARK         '\042'
997    #define CHAR_NUMBER_SIGN            '\043'
998    #define CHAR_DOLLAR_SIGN            '\044'
999    #define CHAR_PERCENT_SIGN           '\045'
1000    #define CHAR_AMPERSAND              '\046'
1001    #define CHAR_APOSTROPHE             '\047'
1002    #define CHAR_LEFT_PARENTHESIS       '\050'
1003    #define CHAR_RIGHT_PARENTHESIS      '\051'
1004    #define CHAR_ASTERISK               '\052'
1005    #define CHAR_PLUS                   '\053'
1006    #define CHAR_COMMA                  '\054'
1007    #define CHAR_MINUS                  '\055'
1008    #define CHAR_DOT                    '\056'
1009    #define CHAR_SLASH                  '\057'
1010    #define CHAR_0                      '\060'
1011    #define CHAR_1                      '\061'
1012    #define CHAR_2                      '\062'
1013    #define CHAR_3                      '\063'
1014    #define CHAR_4                      '\064'
1015    #define CHAR_5                      '\065'
1016    #define CHAR_6                      '\066'
1017    #define CHAR_7                      '\067'
1018    #define CHAR_8                      '\070'
1019    #define CHAR_9                      '\071'
1020    #define CHAR_COLON                  '\072'
1021    #define CHAR_SEMICOLON              '\073'
1022    #define CHAR_LESS_THAN_SIGN         '\074'
1023    #define CHAR_EQUALS_SIGN            '\075'
1024    #define CHAR_GREATER_THAN_SIGN      '\076'
1025    #define CHAR_QUESTION_MARK          '\077'
1026    #define CHAR_COMMERCIAL_AT          '\100'
1027    #define CHAR_A                      '\101'
1028    #define CHAR_B                      '\102'
1029    #define CHAR_C                      '\103'
1030    #define CHAR_D                      '\104'
1031    #define CHAR_E                      '\105'
1032    #define CHAR_F                      '\106'
1033    #define CHAR_G                      '\107'
1034    #define CHAR_H                      '\110'
1035    #define CHAR_I                      '\111'
1036    #define CHAR_J                      '\112'
1037    #define CHAR_K                      '\113'
1038    #define CHAR_L                      '\114'
1039    #define CHAR_M                      '\115'
1040    #define CHAR_N                      '\116'
1041    #define CHAR_O                      '\117'
1042    #define CHAR_P                      '\120'
1043    #define CHAR_Q                      '\121'
1044    #define CHAR_R                      '\122'
1045    #define CHAR_S                      '\123'
1046    #define CHAR_T                      '\124'
1047    #define CHAR_U                      '\125'
1048    #define CHAR_V                      '\126'
1049    #define CHAR_W                      '\127'
1050    #define CHAR_X                      '\130'
1051    #define CHAR_Y                      '\131'
1052    #define CHAR_Z                      '\132'
1053    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1054    #define CHAR_BACKSLASH              '\134'
1055    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1056    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1057    #define CHAR_UNDERSCORE             '\137'
1058    #define CHAR_GRAVE_ACCENT           '\140'
1059    #define CHAR_a                      '\141'
1060    #define CHAR_b                      '\142'
1061    #define CHAR_c                      '\143'
1062    #define CHAR_d                      '\144'
1063    #define CHAR_e                      '\145'
1064    #define CHAR_f                      '\146'
1065    #define CHAR_g                      '\147'
1066    #define CHAR_h                      '\150'
1067    #define CHAR_i                      '\151'
1068    #define CHAR_j                      '\152'
1069    #define CHAR_k                      '\153'
1070    #define CHAR_l                      '\154'
1071    #define CHAR_m                      '\155'
1072    #define CHAR_n                      '\156'
1073    #define CHAR_o                      '\157'
1074    #define CHAR_p                      '\160'
1075    #define CHAR_q                      '\161'
1076    #define CHAR_r                      '\162'
1077    #define CHAR_s                      '\163'
1078    #define CHAR_t                      '\164'
1079    #define CHAR_u                      '\165'
1080    #define CHAR_v                      '\166'
1081    #define CHAR_w                      '\167'
1082    #define CHAR_x                      '\170'
1083    #define CHAR_y                      '\171'
1084    #define CHAR_z                      '\172'
1085    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1086    #define CHAR_VERTICAL_LINE          '\174'
1087    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1088    #define CHAR_TILDE                  '\176'
1089    
1090    #define STR_HT                      "\011"
1091    #define STR_VT                      "\013"
1092    #define STR_FF                      "\014"
1093    #define STR_CR                      "\015"
1094    #define STR_NL                      "\012"
1095    #define STR_BS                      "\010"
1096    #define STR_BEL                     "\007"
1097    #define STR_ESC                     "\033"
1098    #define STR_DEL                     "\177"
1099    
1100    #define STR_SPACE                   "\040"
1101    #define STR_EXCLAMATION_MARK        "\041"
1102    #define STR_QUOTATION_MARK          "\042"
1103    #define STR_NUMBER_SIGN             "\043"
1104    #define STR_DOLLAR_SIGN             "\044"
1105    #define STR_PERCENT_SIGN            "\045"
1106    #define STR_AMPERSAND               "\046"
1107    #define STR_APOSTROPHE              "\047"
1108    #define STR_LEFT_PARENTHESIS        "\050"
1109    #define STR_RIGHT_PARENTHESIS       "\051"
1110    #define STR_ASTERISK                "\052"
1111    #define STR_PLUS                    "\053"
1112    #define STR_COMMA                   "\054"
1113    #define STR_MINUS                   "\055"
1114    #define STR_DOT                     "\056"
1115    #define STR_SLASH                   "\057"
1116    #define STR_0                       "\060"
1117    #define STR_1                       "\061"
1118    #define STR_2                       "\062"
1119    #define STR_3                       "\063"
1120    #define STR_4                       "\064"
1121    #define STR_5                       "\065"
1122    #define STR_6                       "\066"
1123    #define STR_7                       "\067"
1124    #define STR_8                       "\070"
1125    #define STR_9                       "\071"
1126    #define STR_COLON                   "\072"
1127    #define STR_SEMICOLON               "\073"
1128    #define STR_LESS_THAN_SIGN          "\074"
1129    #define STR_EQUALS_SIGN             "\075"
1130    #define STR_GREATER_THAN_SIGN       "\076"
1131    #define STR_QUESTION_MARK           "\077"
1132    #define STR_COMMERCIAL_AT           "\100"
1133    #define STR_A                       "\101"
1134    #define STR_B                       "\102"
1135    #define STR_C                       "\103"
1136    #define STR_D                       "\104"
1137    #define STR_E                       "\105"
1138    #define STR_F                       "\106"
1139    #define STR_G                       "\107"
1140    #define STR_H                       "\110"
1141    #define STR_I                       "\111"
1142    #define STR_J                       "\112"
1143    #define STR_K                       "\113"
1144    #define STR_L                       "\114"
1145    #define STR_M                       "\115"
1146    #define STR_N                       "\116"
1147    #define STR_O                       "\117"
1148    #define STR_P                       "\120"
1149    #define STR_Q                       "\121"
1150    #define STR_R                       "\122"
1151    #define STR_S                       "\123"
1152    #define STR_T                       "\124"
1153    #define STR_U                       "\125"
1154    #define STR_V                       "\126"
1155    #define STR_W                       "\127"
1156    #define STR_X                       "\130"
1157    #define STR_Y                       "\131"
1158    #define STR_Z                       "\132"
1159    #define STR_LEFT_SQUARE_BRACKET     "\133"
1160    #define STR_BACKSLASH               "\134"
1161    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1162    #define STR_CIRCUMFLEX_ACCENT       "\136"
1163    #define STR_UNDERSCORE              "\137"
1164    #define STR_GRAVE_ACCENT            "\140"
1165    #define STR_a                       "\141"
1166    #define STR_b                       "\142"
1167    #define STR_c                       "\143"
1168    #define STR_d                       "\144"
1169    #define STR_e                       "\145"
1170    #define STR_f                       "\146"
1171    #define STR_g                       "\147"
1172    #define STR_h                       "\150"
1173    #define STR_i                       "\151"
1174    #define STR_j                       "\152"
1175    #define STR_k                       "\153"
1176    #define STR_l                       "\154"
1177    #define STR_m                       "\155"
1178    #define STR_n                       "\156"
1179    #define STR_o                       "\157"
1180    #define STR_p                       "\160"
1181    #define STR_q                       "\161"
1182    #define STR_r                       "\162"
1183    #define STR_s                       "\163"
1184    #define STR_t                       "\164"
1185    #define STR_u                       "\165"
1186    #define STR_v                       "\166"
1187    #define STR_w                       "\167"
1188    #define STR_x                       "\170"
1189    #define STR_y                       "\171"
1190    #define STR_z                       "\172"
1191    #define STR_LEFT_CURLY_BRACKET      "\173"
1192    #define STR_VERTICAL_LINE           "\174"
1193    #define STR_RIGHT_CURLY_BRACKET     "\175"
1194    #define STR_TILDE                   "\176"
1195    
1196    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1197    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1198    #define STRING_F0                   STR_F "\0"
1199    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1200    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1201    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1202    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1203    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1204    
1205    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1206    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1207    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1208    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1209    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1210    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1211    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1212    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1213    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1214    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1215    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1216    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1217    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1218    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1219    
1220    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1221    
1222    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1223    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1224    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1225    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1226    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1227    #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
1228    #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
1229    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1230    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1231    #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
1232    
1233    #endif  /* SUPPORT_UTF8 */
1234    
1235  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1236    
1237  #ifndef ESC_e  #ifndef ESC_e
1238  #define ESC_e 27  #define ESC_e CHAR_ESC
1239  #endif  #endif
1240    
1241  #ifndef ESC_f  #ifndef ESC_f
1242  #define ESC_f '\f'  #define ESC_f CHAR_FF
1243  #endif  #endif
1244    
1245  #ifndef ESC_n  #ifndef ESC_n
1246  #define ESC_n '\n'  #define ESC_n CHAR_NL
1247  #endif  #endif
1248    
1249  #ifndef ESC_r  #ifndef ESC_r
1250  #define ESC_r '\r'  #define ESC_r CHAR_CR
1251  #endif  #endif
1252    
1253  /* 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
1254  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1255    
1256  #ifndef ESC_tee  #ifndef ESC_tee
1257  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1258  #endif  #endif
1259    
1260  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1261    
1262  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1263  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1264  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1265  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1266  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1267    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1268    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1269    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1270    #define PT_WORD       8    /* Word - L plus N plus underscore */
1271    
1272  /* 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
1273  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 603  contain UTF-8 characters with values gre Line 1284  contain UTF-8 characters with values gre
1284  /* 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
1285  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
1286  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
1287  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
1288  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
1289  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1290  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1291  ESC_Z to detect the types that may be repeated. These are the types that  
1292  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.
1293  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1294    They must be contiguous, and remain in order so that the replacements can be
1295    looked up from a table.
1296    
1297    The final escape must be ESC_REF as subsequent values are used for
1298    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1299    greater than ESC_b and less than ESC_Z to detect the types that may be
1300    repeated. These are the types that consume characters. If any new escapes are
1301    put in between that don't consume a character, that code will have to change.
1302    */
1303    
1304  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,
1305         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,
1306         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,
1307           ESC_E, ESC_Q, ESC_g, ESC_k,
1308           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1309           ESC_REF };
1310    
1311  /* 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
1312  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.
1313    
1314  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1315  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
1316  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1317    
1318  enum {  enum {
1319    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 639  enum { Line 1331  enum {
1331    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1332    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1333    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 */  
1334    
1335    /* These three must follow the previous three, in the same order. There's a    OP_ANY,            /* 12 Match any character except newline */
1336    check for >= SBRA to distinguish the two sets. */    OP_ALLANY,         /* 13 Match any character */
1337      OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1338      OP_NOTPROP,        /* 15 \P (not Unicode property) */
1339      OP_PROP,           /* 16 \p (Unicode property) */
1340      OP_ANYNL,          /* 17 \R (any newline sequence) */
1341      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1342      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1343      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1344      OP_VSPACE,         /* 21 \v (vertical whitespace) */
1345      OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1346      OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1347      OP_EOD,            /* 24 End of data: \z */
1348    
1349      OP_CIRC,           /* 25 Start of line - not multiline */
1350      OP_CIRCM,          /* 26 Start of line - multiline */
1351      OP_DOLL,           /* 27 End of line - not multiline */
1352      OP_DOLLM,          /* 28 End of line - multiline */
1353      OP_CHAR,           /* 29 Match one character, casefully */
1354      OP_CHARI,          /* 30 Match one character, caselessly */
1355      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1356      OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1357    
1358      /* The following sets of 13 opcodes must always be kept in step because
1359      the offset from the first one is used to generate the others. */
1360    
1361      /**** Single characters, caseful, must precede the caseless ones ****/
1362    
1363      OP_STAR,           /* 33 The maximizing and minimizing versions of */
1364      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1365      OP_PLUS,           /* 35 the minimizing one second. */
1366      OP_MINPLUS,        /* 36 */
1367      OP_QUERY,          /* 37 */
1368      OP_MINQUERY,       /* 38 */
1369    
1370      OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1371      OP_MINUPTO,        /* 40 */
1372      OP_EXACT,          /* 41 Exactly n matches */
1373    
1374      OP_POSSTAR,        /* 42 Possessified star, caseful */
1375      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1376      OP_POSQUERY,       /* 44 Posesssified query, caseful */
1377      OP_POSUPTO,        /* 45 Possessified upto, caseful */
1378    
1379      /**** Single characters, caseless, must follow the caseful ones */
1380    
1381      OP_STARI,          /* 46 */
1382      OP_MINSTARI,       /* 47 */
1383      OP_PLUSI,          /* 48 */
1384      OP_MINPLUSI,       /* 49 */
1385      OP_QUERYI,         /* 50 */
1386      OP_MINQUERYI,      /* 51 */
1387    
1388      OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1389      OP_MINUPTOI,       /* 53 */
1390      OP_EXACTI,         /* 54 */
1391    
1392      OP_POSSTARI,       /* 55 Possessified star, caseless */
1393      OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1394      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1395      OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1396    
1397      /**** The negated ones must follow the non-negated ones, and match them ****/
1398      /**** Negated single character, caseful; must precede the caseless ones ****/
1399    
1400      OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1401      OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1402      OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1403      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1404      OP_NOTQUERY,       /* 63 */
1405      OP_NOTMINQUERY,    /* 64 */
1406    
1407      OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1408      OP_NOTMINUPTO,     /* 66 */
1409      OP_NOTEXACT,       /* 67 Exactly n matches */
1410    
1411      OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1412      OP_NOTPOSPLUS,     /* 69 */
1413      OP_NOTPOSQUERY,    /* 70 */
1414      OP_NOTPOSUPTO,     /* 71 */
1415    
1416      /**** Negated single character, caseless; must follow the caseful ones ****/
1417    
1418      OP_NOTSTARI,       /* 72 */
1419      OP_NOTMINSTARI,    /* 73 */
1420      OP_NOTPLUSI,       /* 74 */
1421      OP_NOTMINPLUSI,    /* 75 */
1422      OP_NOTQUERYI,      /* 76 */
1423      OP_NOTMINQUERYI,   /* 77 */
1424    
1425      OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1426      OP_NOTMINUPTOI,    /* 79 */
1427      OP_NOTEXACTI,      /* 80 Exactly n matches */
1428    
1429      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1430      OP_NOTPOSPLUSI,    /* 82 */
1431      OP_NOTPOSQUERYI,   /* 83 */
1432      OP_NOTPOSUPTOI,    /* 84 */
1433    
1434      /**** Character types ****/
1435    
1436      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1437      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1438      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1439      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1440      OP_TYPEQUERY,      /* 89 */
1441      OP_TYPEMINQUERY,   /* 90 */
1442    
1443      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1444      OP_TYPEMINUPTO,    /* 92 */
1445      OP_TYPEEXACT,      /* 93 Exactly n matches */
1446    
1447      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1448      OP_TYPEPOSPLUS,    /* 95 */
1449      OP_TYPEPOSQUERY,   /* 96 */
1450      OP_TYPEPOSUPTO,    /* 97 */
1451    
1452      /* These are used for character classes and back references; only the
1453      first six are the same as the sets above. */
1454    
1455      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1456      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1457      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1458      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1459      OP_CRQUERY,        /* 102 */
1460      OP_CRMINQUERY,     /* 103 */
1461    
1462      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1463      OP_CRMINRANGE,     /* 105 */
1464    
1465      /* End of quantifier opcodes */
1466    
1467      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1468      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1469                                  class - the difference is relevant only when a
1470                                  UTF-8 character > 255 is encountered. */
1471      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1472                                  class. This does both positive and negative. */
1473      OP_REF,            /* 109 Match a back reference, casefully */
1474      OP_REFI,           /* 110 Match a back reference, caselessly */
1475      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1476      OP_CALLOUT,        /* 112 Call out to external function if provided */
1477    
1478      OP_ALT,            /* 113 Start of alternation */
1479      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1480      OP_KETRMAX,        /* 115 These two must remain together and in this */
1481      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1482      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1483    
1484      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1485      asserts must remain in order. */
1486    
1487      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1488      OP_ASSERT,         /* 119 Positive lookahead */
1489      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1490      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1491      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1492    
1493      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1494      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1495      subpattern that isn't an assertion. The POS versions must immediately follow
1496      the non-POS versions in each case. */
1497    
1498      OP_ONCE,           /* 123 Atomic group, contains captures */
1499      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1500      OP_BRA,            /* 125 Start of non-capturing bracket */
1501      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1502      OP_CBRA,           /* 127 Start of capturing bracket */
1503      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1504      OP_COND,           /* 129 Conditional group */
1505    
1506    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    /* These five must follow the previous five, in the same order. There's a
1507    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 */  
1508    
1509    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1510    OP_BRAMINZERO,     /* 103 order. */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1511      OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1512      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1513      OP_SCOND,          /* 134 Conditional group, check empty */
1514    
1515      /* The next two pairs must (respectively) be kept together. */
1516    
1517      OP_CREF,           /* 135 Used to hold a capture number as condition */
1518      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1519      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1520      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1521      OP_DEF,            /* 139 The DEFINE condition */
1522    
1523      OP_BRAZERO,        /* 140 These two must remain together and in this */
1524      OP_BRAMINZERO,     /* 141 order. */
1525      OP_BRAPOSZERO,     /* 142 */
1526    
1527    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1528    
1529    OP_PRUNE,          /* 104 */    OP_MARK,           /* 143 always has an argument */
1530    OP_SKIP,           /* 105 */    OP_PRUNE,          /* 144 */
1531    OP_THEN,           /* 106 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1532    OP_COMMIT,         /* 107 */    OP_SKIP,           /* 146 */
1533      OP_SKIP_ARG,       /* 147 same, but with argument */
1534      OP_THEN,           /* 148 */
1535      OP_THEN_ARG,       /* 149 same, but with argument */
1536      OP_COMMIT,         /* 150 */
1537    
1538    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1539    
1540    OP_FAIL,           /* 108 */    OP_FAIL,           /* 151 */
1541    OP_ACCEPT,         /* 109 */    OP_ACCEPT,         /* 152 */
1542      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1543      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1544    
1545    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1546    
1547    OP_SKIPZERO        /* 110 */    OP_SKIPZERO,       /* 155 */
1548    
1549      /* This is not an opcode, but is used to check that tables indexed by opcode
1550      are the correct length, in order to catch updating errors - there have been
1551      some in the past. */
1552    
1553      OP_TABLE_LENGTH
1554  };  };
1555    
1556    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1557    definitions that follow must also be updated to match. There are also tables
1558    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1559    
1560    
1561  /* 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
1562  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
1563    only in pcre_printint.c, which fills out the full names in many cases (and in
1564    some cases doesn't actually use these names at all). */
1565    
1566  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1567    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1568    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1569    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1570    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1571    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1572    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1573      "{", "{", "{",                                                  \
1574    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1575    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1576      "{", "{", "{",                                                  \
1577      "*+","++", "?+", "{",                                           \
1578      "*", "*?", "+", "+?", "?", "??",                                \
1579      "{", "{", "{",                                                  \
1580      "*+","++", "?+", "{",                                           \
1581      "*", "*?", "+", "+?", "?", "??",                                \
1582      "{", "{", "{",                                                  \
1583    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1584    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1585    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1586    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1587    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1588    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1589    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1590    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1591    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once", "Once_NC",                                              \
1592    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Bra", "BraPos", "CBra", "CBraPos",                             \
1593    "Skip zero"    "Cond",                                                         \
1594      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1595      "SCond",                                                        \
1596      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1597      "Brazero", "Braminzero", "Braposzero",                          \
1598      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1599      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1600      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1601      "Close", "Skip zero"
1602    
1603    
1604  /* 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 1614  in UTF-8 mode. The code that uses this t
1614    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1615    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1616    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1617    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1618    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1619    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1620    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1621      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1622    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1623    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1624    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1625    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1626    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1627    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1628      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1629    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1630      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1631      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1632      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1633    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1634    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1635    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1636    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1637      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1638      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1639      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1640    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1641    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1642    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 845  in UTF-8 mode. The code that uses this t Line 1648  in UTF-8 mode. The code that uses this t
1648   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1649    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1650    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1651      3,                             /* REFI                                   */ \
1652    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1653    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1654    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1655    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1656    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1657    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1658      1+LINK_SIZE,                   /* KetRpos                                */ \
1659      1+LINK_SIZE,                   /* Reverse                                */ \
1660    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1661    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1662    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1663    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1664    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1665      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1666    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1667      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1668    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1669      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1670    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1671    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1672      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1673    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1674      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1675    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1676    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1677    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1678    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1679    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1680    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1681    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1682      1, 3,                          /* THEN, THEN_ARG                         */ \
1683      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1684      3, 1                           /* CLOSE, SKIPZERO  */
1685    
1686  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1687    condition. */
1688    
1689  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1690    
1691  /* 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
1692  tracked. */  easily be tracked. When a new number is added, the table called eint in
1693    pcreposix.c must be updated. */
1694    
1695  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1696         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 1698  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1698         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1699         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1700         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1701         ERR60, ERR61, ERR62, ERR63 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1702           ERR70, ERRCOUNT };
1703    
1704  /* 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
1705  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 1715  Because people can now save and re-use c
1715  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
1716  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
1717  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1718  NOTE NOTE NOTE:  NOTE NOTE NOTE
1719  */  */
1720    
1721  typedef struct real_pcre {  typedef struct real_pcre {
# Line 918  typedef struct real_pcre { Line 1733  typedef struct real_pcre {
1733    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1734    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1735    
1736    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1737    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1738  } real_pcre;  } real_pcre;
1739    
1740  /* 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 1742  remark (see NOTE above) about extending
1742    
1743  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1744    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1745    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1746    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
1747      pcre_uint32 minlength;          /* Minimum subject length */
1748  } pcre_study_data;  } pcre_study_data;
1749    
1750    /* Structure for building a chain of open capturing subpatterns during
1751    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1752    encountered. This is also used to identify subpatterns that contain recursive
1753    back references to themselves, so that they can be made atomic. */
1754    
1755    typedef struct open_capitem {
1756      struct open_capitem *next;    /* Chain link */
1757      pcre_uint16 number;           /* Capture number */
1758      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1759    } open_capitem;
1760    
1761  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1762  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1763    
1764  typedef struct compile_data {  typedef struct compile_data {
1765    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1766    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1767    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1768    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1769    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1770    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1771    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1772    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1773    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
1774    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
1775    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
1776    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
1777    int  bracount;                /* Count of capturing parens as we compile */    int  name_entry_size;             /* Size of each entry */
1778    int  final_bracount;          /* Saved value after first pass */    int  bracount;                    /* Count of capturing parens as we compile */
1779    int  top_backref;             /* Maximum back reference */    int  final_bracount;              /* Saved value after first pass */
1780    unsigned int backref_map;     /* Bitmap of low back refs */    int  top_backref;                 /* Maximum back reference */
1781    int  external_options;        /* External (initial) options */    unsigned int backref_map;         /* Bitmap of low back refs */
1782    int  external_flags;          /* External flag bits to be set */    int  assert_depth;                /* Depth of nested assertions */
1783    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_options;            /* External (initial) options */
1784    BOOL had_accept;              /* (*ACCEPT) encountered */    int  external_flags;              /* External flag bits to be set */
1785    int  nltype;                  /* Newline type */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1786    int  nllen;                   /* Newline string length */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1787    uschar nl[4];                 /* Newline string when fixed length */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1788      int  nltype;                      /* Newline type */
1789      int  nllen;                       /* Newline string length */
1790      pcre_uchar nl[4];                 /* Newline string when fixed length */
1791  } compile_data;  } compile_data;
1792    
1793  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1794  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1795    
1796  typedef struct branch_chain {  typedef struct branch_chain {
1797    struct branch_chain *outer;    struct branch_chain *outer;
1798    uschar *current;    pcre_uchar *current_branch;
1799  } branch_chain;  } branch_chain;
1800    
1801  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1802  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1803    
1804  typedef struct recursion_info {  typedef struct recursion_info {
1805    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1806    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1807    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1808    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
1809    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 */  
1810  } recursion_info;  } recursion_info;
1811    
1812    /* A similar structure for pcre_dfa_exec(). */
1813    
1814    typedef struct dfa_recursion_info {
1815      struct dfa_recursion_info *prevrec;
1816      int group_num;
1817      PCRE_PUCHAR subject_position;
1818    } dfa_recursion_info;
1819    
1820  /* 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
1821  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
1822  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1823    pcre_exec(). */
1824    
1825  typedef struct eptrblock {  typedef struct eptrblock {
1826    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
1827    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
1828  } eptrblock;  } eptrblock;
1829    
1830    
# Line 997  typedef struct match_data { Line 1835  typedef struct match_data {
1835    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
1836    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
1837    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
1838    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
1839    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
1840    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
1841    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
1842    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
1843    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
1844    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
1845    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
1846    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
1847    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
1848    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
1849    BOOL   utf8;                  /* UTF8 flag */    BOOL   offset_overflow;         /* Set if too many extractions */
1850    BOOL   endonly;               /* Dollar not before final \n */    BOOL   notbol;                  /* NOTBOL flag */
1851    BOOL   notempty;              /* Empty string match not wanted */    BOOL   noteol;                  /* NOTEOL flag */
1852    BOOL   partial;               /* PARTIAL flag */    BOOL   utf8;                    /* UTF8 flag */
1853    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
1854    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   use_ucp;                 /* PCRE_UCP flag */
1855    const uschar *start_code;     /* For use when recursing */    BOOL   endonly;                 /* Dollar not before final \n */
1856    USPTR  start_subject;         /* Start of the subject string */    BOOL   notempty;                /* Empty string match not wanted */
1857    USPTR  end_subject;           /* End of the subject string */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
1858    USPTR  start_match_ptr;       /* Start of matched string */    BOOL   hitend;                  /* Hit the end of the subject at some point */
1859    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
1860    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
1861    int    capture_last;          /* Most recent capture number */    const  pcre_uchar *start_code;  /* For use when recursing */
1862    int    start_offset;          /* The start offset value */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
1863    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    PCRE_PUCHAR end_subject;        /* End of the subject string */
1864    int    eptrn;                 /* Next free eptrblock */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
1865    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
1866    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
1867      int    partial;                 /* PARTIAL options */
1868      int    end_offset_top;          /* Highwater mark at end of match */
1869      int    capture_last;            /* Most recent capture number */
1870      int    start_offset;            /* The start offset value */
1871      int    match_function_type;     /* Set for certain special calls of MATCH() */
1872      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
1873      int    eptrn;                   /* Next free eptrblock */
1874      recursion_info *recursive;      /* Linked list of recursion data */
1875      void  *callout_data;            /* To pass back to callouts */
1876      const  pcre_uchar *mark;        /* Mark pointer to pass back */
1877      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
1878  } match_data;  } match_data;
1879    
1880  /* 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
1881  functions. */  functions. */
1882    
1883  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1884    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
1885    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
1886    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
1887    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
1888    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
1889    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
1890    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
1891    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
1892    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
1893    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
1894      pcre_uchar nl[4];                 /* Newline string when fixed */
1895      void *callout_data;               /* To pass back to callouts */
1896      dfa_recursion_info *recursive;    /* Linked list of recursion data */
1897  } dfa_match_data;  } dfa_match_data;
1898    
1899  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1094  of the exported public functions. They h Line 1946  of the exported public functions. They h
1946  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
1947  pcre_tables.c module. */  pcre_tables.c module. */
1948    
1949  extern const int    _pcre_utf8_table1[];  extern const int            _pcre_utf8_table1[];
1950  extern const int    _pcre_utf8_table2[];  extern const int            _pcre_utf8_table2[];
1951  extern const int    _pcre_utf8_table3[];  extern const int            _pcre_utf8_table3[];
1952  extern const uschar _pcre_utf8_table4[];  extern const pcre_uint8     _pcre_utf8_table4[];
1953    
1954  extern const int    _pcre_utf8_table1_size;  extern const int            _pcre_utf8_table1_size;
1955    
1956  extern const char   _pcre_utt_names[];  extern const char           _pcre_utt_names[];
1957  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1958  extern const int _pcre_utt_size;  extern const int            _pcre_utt_size;
1959    
1960  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     _pcre_default_tables[];
1961    
1962  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     _pcre_OP_lengths[];
1963    
1964    
1965  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
1966  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
1967  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1968    
1969  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  /* String comparison functions. */
1970                        int *, BOOL);  #ifdef COMPILE_PCRE8
1971  extern int          _pcre_ord2utf8(int, uschar *);  
1972  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
1973                        const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
1974  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
1975  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
1976  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
1977  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,    strncmp((char *)(str1), (char *)(str2), (num))
1978                        int *, BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
1979  extern BOOL         _pcre_xclass(int, const uschar *);    strncmp((char *)(str1), (str2), (num))
1980    #define STRLEN_UC(str) strlen(str)
1981    
1982    #else
1983    
1984    extern int               _pcre_strcmp_uc_uc(const pcre_uchar *,
1985                               const pcre_uchar *);
1986    extern int               _pcre_strcmp_uc_c8(const pcre_uchar *,
1987                               const char *);
1988    extern int               _pcre_strncmp_uc_uc(const pcre_uchar *,
1989                               const pcre_uchar *, unsigned int num);
1990    extern int               _pcre_strncmp_uc_c8(const pcre_uchar *,
1991                               const char *, unsigned int num);
1992    extern unsigned int      _pcre_strlen_uc(const pcre_uchar *str);
1993    
1994    #define STRCMP_UC_UC(str1, str2) \
1995      _pcre_strcmp_uc_uc((str1), (str2))
1996    #define STRCMP_UC_C8(str1, str2) \
1997      _pcre_strcmp_uc_c8((str1), (str2))
1998    #define STRNCMP_UC_UC(str1, str2, num) \
1999      _pcre_strncmp_uc_uc((str1), (str2), (num))
2000    #define STRNCMP_UC_C8(str1, str2, num) \
2001      _pcre_strncmp_uc_c8((str1), (str2), (num))
2002    #define STRLEN_UC(str) _pcre_strlen_uc(str)
2003    
2004    #endif /* COMPILE_PCRE8 */
2005    
2006    extern const pcre_uchar *_pcre_find_bracket(const pcre_uchar *, BOOL, int);
2007    extern BOOL              _pcre_is_newline(PCRE_PUCHAR, int, PCRE_PUCHAR,
2008                               int *, BOOL);
2009    extern int               _pcre_ord2utf8(int, pcre_uint8 *);
2010    extern real_pcre        *_pcre_try_flipped(const real_pcre *, real_pcre *,
2011                               const pcre_study_data *, pcre_study_data *);
2012    #ifndef COMPILE_PCRE16
2013    extern int               _pcre_valid_utf8(PCRE_PUCHAR, int, int *);
2014    #else
2015    extern int               _pcre16_valid_utf16(PCRE_PUCHAR, int, int *);
2016    #endif
2017    extern BOOL              _pcre_was_newline(PCRE_PUCHAR, int, PCRE_PUCHAR,
2018                               int *, BOOL);
2019    extern BOOL              _pcre_xclass(int, const pcre_uchar *);
2020    
2021    #ifdef SUPPORT_JIT
2022    extern void              _pcre_jit_compile(const real_pcre *, pcre_extra *);
2023    extern int               _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,
2024                               int, int, int, int, int *, int);
2025    extern void              _pcre_jit_free(void *);
2026    #endif
2027    
2028    /* Unicode character database (UCD) */
2029    
2030    typedef struct {
2031      pcre_uint8 script;
2032      pcre_uint8 chartype;
2033      pcre_int32 other_case;
2034    } ucd_record;
2035    
2036    extern const ucd_record  _pcre_ucd_records[];
2037    extern const pcre_uint8  _pcre_ucd_stage1[];
2038    extern const pcre_uint16 _pcre_ucd_stage2[];
2039    extern const int         _pcre_ucp_gentype[];
2040    #ifdef SUPPORT_JIT
2041    extern const int         _pcre_ucp_typerange[];
2042    #endif
2043    
2044    /* UCD access macros */
2045    
2046    #define UCD_BLOCK_SIZE 128
2047    #define GET_UCD(ch) (_pcre_ucd_records + \
2048            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
2049            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2050    
2051    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2052    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2053    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
2054    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2055    
2056  #endif  #endif
2057    

Legend:
Removed from v.335  
changed lines
  Added in v.763

  ViewVC Help
Powered by ViewVC 1.1.5