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

Diff of /code/trunk/pcre_internal.h

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

revision 1128 by chpe, Thu Oct 18 18:35:08 2012 UTC revision 1425 by ph10, Tue Dec 31 17:44:40 2013 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) 1997-2012 University of Cambridge             Copyright (c) 1997-2013 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 53  depending on the PRIV macro. */ Line 53  depending on the PRIV macro. */
53  #endif  #endif
54    
55  /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */  /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56    
57  #if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32  #if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32
58  #define COMPILE_PCRE8  #define COMPILE_PCRE8
59  #endif  #endif
# Line 111  setjmp and stdarg are used is when NO_RE Line 112  setjmp and stdarg are used is when NO_RE
112  #include <stdlib.h>  #include <stdlib.h>
113  #include <string.h>  #include <string.h>
114    
115    /* Valgrind (memcheck) support */
116    
117    #ifdef SUPPORT_VALGRIND
118    #include <valgrind/memcheck.h>
119    #endif
120    
121  /* When compiling a DLL for Windows, the exported symbols have to be declared  /* When compiling a DLL for Windows, the exported symbols have to be declared
122  using some MS magic. I found some useful information on this web page:  using some MS magic. I found some useful information on this web page:
123  http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the  http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
# Line 187  preprocessor time in standard C environm Line 194  preprocessor time in standard C environm
194  typedef unsigned char pcre_uint8;  typedef unsigned char pcre_uint8;
195    
196  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
197    typedef unsigned short pcre_uint16;  typedef unsigned short pcre_uint16;
198    typedef short pcre_int16;  typedef short pcre_int16;
199    #define PCRE_UINT16_MAX USHRT_MAX
200    #define PCRE_INT16_MAX SHRT_MAX
201  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
202    typedef unsigned int pcre_uint16;  typedef unsigned int pcre_uint16;
203    typedef int pcre_int16;  typedef int pcre_int16;
204    #define PCRE_UINT16_MAX UINT_MAX
205    #define PCRE_INT16_MAX INT_MAX
206  #else  #else
207    #error Cannot determine a type for 16-bit unsigned integers  #error Cannot determine a type for 16-bit integers
208  #endif  #endif
209    
210  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295U
211    typedef unsigned int pcre_uint32;  typedef unsigned int pcre_uint32;
212    typedef int pcre_int32;  typedef int pcre_int32;
213  #elif ULONG_MAX == 4294967295  #define PCRE_UINT32_MAX UINT_MAX
214    typedef unsigned long int pcre_uint32;  #define PCRE_INT32_MAX INT_MAX
215    typedef long int pcre_int32;  #elif ULONG_MAX == 4294967295UL
216    typedef unsigned long int pcre_uint32;
217    typedef long int pcre_int32;
218    #define PCRE_UINT32_MAX ULONG_MAX
219    #define PCRE_INT32_MAX LONG_MAX
220  #else  #else
221    #error Cannot determine a type for 32-bit unsigned integers  #error Cannot determine a type for 32-bit integers
222  #endif  #endif
223    
224  /* When checking for integer overflow in pcre_compile(), we need to handle  /* When checking for integer overflow in pcre_compile(), we need to handle
# Line 214  stdint.h is available, include it; it ma Line 229  stdint.h is available, include it; it ma
229  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
230  by "configure". */  by "configure". */
231    
232  #if HAVE_STDINT_H  #if defined HAVE_STDINT_H
233  #include <stdint.h>  #include <stdint.h>
234  #elif HAVE_INTTYPES_H  #elif defined HAVE_INTTYPES_H
235  #include <inttypes.h>  #include <inttypes.h>
236  #endif  #endif
237    
# Line 272  typedef pcre_uint32 pcre_uchar; Line 287  typedef pcre_uint32 pcre_uchar;
287  #define MAX_255(c) ((c) <= 255u)  #define MAX_255(c) ((c) <= 255u)
288  #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))  #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
289    
 /* Assert that pcre_uchar32 is a 32-bit type */  
 typedef int __assert_pcre_uchar32_size[sizeof(pcre_uchar) == 4 ? 1 : -1];  
   
290  #else  #else
291  #error Unsupported compiling mode  #error Unsupported compiling mode
292  #endif /* COMPILE_PCRE[8|16|32] */  #endif /* COMPILE_PCRE[8|16|32] */
# Line 304  start/end of string field names are. */ Line 316  start/end of string field names are. */
316         &(NLBLOCK->nllen), utf)) \         &(NLBLOCK->nllen), utf)) \
317      : \      : \
318      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
319       RAWUCHARTEST(p) == NLBLOCK->nl[0] && \       *p == NLBLOCK->nl[0] && (NLBLOCK->nllen == 1 || p[1] == NLBLOCK->nl[1]) \
      (NLBLOCK->nllen == 1 || RAWUCHARTEST(p+1) == NLBLOCK->nl[1])       \  
320      ) \      ) \
321    )    )
322    
# Line 318  start/end of string field names are. */ Line 329  start/end of string field names are. */
329         &(NLBLOCK->nllen), utf)) \         &(NLBLOCK->nllen), utf)) \
330      : \      : \
331      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
332       RAWUCHARTEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] &&              \       *(p - NLBLOCK->nllen) == NLBLOCK->nl[0] &&              \
333       (NLBLOCK->nllen == 1 || RAWUCHARTEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \       (NLBLOCK->nllen == 1 || *(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \
334      ) \      ) \
335    )    )
336    
# Line 344  values. */ Line 355  values. */
355  #include "pcre.h"  #include "pcre.h"
356  #include "ucp.h"  #include "ucp.h"
357    
358    #ifdef COMPILE_PCRE32
359    /* Assert that the public PCRE_UCHAR32 is a 32-bit type */
360    typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1];
361    #endif
362    
363  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
364  need to have their names changed. PCRE must be compiled with the -DVPCOMPAT  need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
365  option on the command line. */  option on the command line. */
# Line 523  capturing parenthesis numbers in back re Line 539  capturing parenthesis numbers in back re
539    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
540    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
541    
542    /* For reasons that I do not understand, the expression in this GET2 macro is
543    treated by gcc as a signed expression, even when a is declared as unsigned. It
544    seems that any kind of arithmetic results in a signed value. */
545    
546  #define GET2(a,n) \  #define GET2(a,n) \
547    (((a)[n] << 8) | (a)[(n)+1])    (unsigned int)(((a)[n] << 8) | (a)[(n)+1])
548    
549  #elif defined COMPILE_PCRE16  #elif defined COMPILE_PCRE16
550    
# Line 579  we don't even define them. */ Line 599  we don't even define them. */
599  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
600  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
601  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
 #define RAWUCHAR(eptr) (*(eptr))  
 #define RAWUCHARINC(eptr) (*(eptr)++)  
 #define RAWUCHARTEST(eptr) (*(eptr))  
 #define RAWUCHARINCTEST(eptr) (*(eptr)++)  
602  /* #define GETCHARLENTEST(c, eptr, len) */  /* #define GETCHARLENTEST(c, eptr, len) */
603  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
604  /* #define FORWARDCHAR(eptr) */  /* #define FORWARDCHAR(eptr) */
# Line 755  do not know if we are in UTF-8 mode. */ Line 771  do not know if we are in UTF-8 mode. */
771    c = *eptr; \    c = *eptr; \
772    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
773    
 /* Returns the next uchar, not advancing the pointer. This is called when  
 we know we are in UTF mode. */  
   
 #define RAWUCHAR(eptr) \  
   (*(eptr))  
   
 /* Returns the next uchar, advancing the pointer. This is called when  
 we know we are in UTF mode. */  
   
 #define RAWUCHARINC(eptr) \  
   (*(eptr)++)  
   
 /* Returns the next uchar, testing for UTF mode, and not advancing the  
 pointer. */  
   
 #define RAWUCHARTEST(eptr) \  
   (*(eptr))  
   
 /* Returns the next uchar, testing for UTF mode, advancing the  
 pointer. */  
   
 #define RAWUCHARINCTEST(eptr) \  
   (*(eptr)++)  
   
774  /* 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
775  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
776  because almost all calls are already within a block of UTF-8 only code. */  because almost all calls are already within a block of UTF-8 only code. */
# Line 874  we do not know if we are in UTF-16 mode. Line 866  we do not know if we are in UTF-16 mode.
866    c = *eptr; \    c = *eptr; \
867    if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);    if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
868    
 /* Returns the next uchar, not advancing the pointer. This is called when  
 we know we are in UTF mode. */  
   
 #define RAWUCHAR(eptr) \  
   (*(eptr))  
   
 /* Returns the next uchar, advancing the pointer. This is called when  
 we know we are in UTF mode. */  
   
 #define RAWUCHARINC(eptr) \  
   (*(eptr)++)  
   
 /* Returns the next uchar, testing for UTF mode, and not advancing the  
 pointer. */  
   
 #define RAWUCHARTEST(eptr) \  
   (*(eptr))  
   
 /* Returns the next uchar, testing for UTF mode, advancing the  
 pointer. */  
   
 #define RAWUCHARINCTEST(eptr) \  
   (*(eptr)++)  
   
869  /* 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
870  it is. This is called only in UTF-16 mode - we don't put a test within the  it is. This is called only in UTF-16 mode - we don't put a test within the
871  macro because almost all calls are already within a block of UTF-16 only  macro because almost all calls are already within a block of UTF-16 only
# Line 921  into one pcre_uchar unit. */ Line 889  into one pcre_uchar unit. */
889  #define GET_EXTRALEN(c) (0)  #define GET_EXTRALEN(c) (0)
890  #define NOT_FIRSTCHAR(c) (0)  #define NOT_FIRSTCHAR(c) (0)
891    
 #define UTF32_MASK (0x1fffffu)  
   
892  /* Get the next UTF-32 character, not advancing the pointer. This is called when  /* Get the next UTF-32 character, not advancing the pointer. This is called when
893  we know we are in UTF-32 mode. */  we know we are in UTF-32 mode. */
894    
895  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
896    c = (*eptr) & UTF32_MASK;    c = *(eptr);
897    
898  /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the  /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the
899  pointer. */  pointer. */
900    
901  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
902    c = *eptr; \    c = *(eptr);
   if (utf) c &= UTF32_MASK;  
903    
904  /* Get the next UTF-32 character, advancing the pointer. This is called when we  /* Get the next UTF-32 character, advancing the pointer. This is called when we
905  know we are in UTF-32 mode. */  know we are in UTF-32 mode. */
906    
907  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
908    c = (*eptr++) & UTF32_MASK;    c = *((eptr)++);
909    
910  /* Get the next character, testing for UTF-32 mode, and advancing the pointer.  /* Get the next character, testing for UTF-32 mode, and advancing the pointer.
911  This is called when we don't know if we are in UTF-32 mode. */  This is called when we don't know if we are in UTF-32 mode. */
912    
913  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
914    c = *eptr++; \    c = *((eptr)++);
   if (utf) c &= UTF32_MASK;  
915    
916  /* Get the next UTF-32 character, not advancing the pointer, not incrementing  /* Get the next UTF-32 character, not advancing the pointer, not incrementing
917  length (since all UTF-32 is of length 1). This is called when we know we are in  length (since all UTF-32 is of length 1). This is called when we know we are in
# Line 963  This is called when we do not know if we Line 927  This is called when we do not know if we
927  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
928    GETCHARTEST(c, eptr)    GETCHARTEST(c, eptr)
929    
 /* Returns the next uchar, not advancing the pointer. This is called when  
 we know we are in UTF mode. */  
   
 #define RAWUCHAR(eptr) \  
   (*(eptr) & UTF32_MASK)  
   
 /* Returns the next uchar, advancing the pointer. This is called when  
 we know we are in UTF mode. */  
   
 #define RAWUCHARINC(eptr) \  
   (*(eptr)++ & UTF32_MASK)  
   
 /* Returns the next uchar, testing for UTF mode, and not advancing the  
 pointer. */  
   
 #define RAWUCHARTEST(eptr) \  
   (utf ? (*(eptr) & UTF32_MASK) : *(eptr))  
   
 /* Returns the next uchar, testing for UTF mode, advancing the  
 pointer. */  
   
 #define RAWUCHARINCTEST(eptr) \  
   (utf ? (*(eptr)++ & UTF32_MASK) : *(eptr)++)  
   
930  /* 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
931  it is. This is called only in UTF-32 mode - we don't put a test within the  it is. This is called only in UTF-32 mode - we don't put a test within the
932  macro because almost all calls are already within a block of UTF-32 only  macro because almost all calls are already within a block of UTF-32 only
933  code.  code.
934  These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */  These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */
935    
936  #define BACKCHAR(eptr) do { } while (0)  #define BACKCHAR(eptr) do { } while (0)
# Line 1017  are defined. Line 957  are defined.
957  These values are also required as lists in pcre_compile.c when processing \h,  These values are also required as lists in pcre_compile.c when processing \h,
958  \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but  \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but
959  macros that define the values are here so that all the definitions are  macros that define the values are here so that all the definitions are
960  together. The lists must be in ascending character order, terminated by  together. The lists must be in ascending character order, terminated by
961  NOTACHAR (which is 0xffffffff).  NOTACHAR (which is 0xffffffff).
962    
963  Any changes should ensure that the various macros are kept in step with each  Any changes should ensure that the various macros are kept in step with each
# Line 1031  other. NOTE: The values also appear in p Line 971  other. NOTE: The values also appear in p
971    CHAR_HT, CHAR_SPACE, 0xa0, \    CHAR_HT, CHAR_SPACE, 0xa0, \
972    0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \    0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \
973    0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \    0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \
974    NOTACHAR    NOTACHAR
975    
976  #define HSPACE_MULTIBYTE_CASES \  #define HSPACE_MULTIBYTE_CASES \
977    case 0x1680:  /* OGHAM SPACE MARK */ \    case 0x1680:  /* OGHAM SPACE MARK */ \
# Line 1055  other. NOTE: The values also appear in p Line 995  other. NOTE: The values also appear in p
995    case CHAR_HT: \    case CHAR_HT: \
996    case CHAR_SPACE: \    case CHAR_SPACE: \
997    case 0xa0     /* NBSP */    case 0xa0     /* NBSP */
998    
999  #define HSPACE_CASES \  #define HSPACE_CASES \
1000    HSPACE_BYTE_CASES: \    HSPACE_BYTE_CASES: \
1001    HSPACE_MULTIBYTE_CASES    HSPACE_MULTIBYTE_CASES
1002    
1003  #define VSPACE_LIST \  #define VSPACE_LIST \
1004    CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR    CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR
1005    
1006  #define VSPACE_MULTIBYTE_CASES \  #define VSPACE_MULTIBYTE_CASES \
1007    case 0x2028:    /* LINE SEPARATOR */ \    case 0x2028:    /* LINE SEPARATOR */ \
# Line 1091  other. NOTE: The values also appear in p Line 1031  other. NOTE: The values also appear in p
1031    
1032  #ifdef EBCDIC_NL25  #ifdef EBCDIC_NL25
1033  #define VSPACE_LIST \  #define VSPACE_LIST \
1034    CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR    CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR
1035  #else  #else
1036  #define VSPACE_LIST \  #define VSPACE_LIST \
1037    CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR    CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR
1038  #endif  #endif
1039    
1040  #define VSPACE_BYTE_CASES \  #define VSPACE_BYTE_CASES \
1041    case CHAR_LF: \    case CHAR_LF: \
# Line 1112  other. NOTE: The values also appear in p Line 1052  other. NOTE: The values also appear in p
1052    
1053    
1054  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
1055  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 they were
1056  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as  moved to a 16-bit flags word - which got almost full, so now they are in a
1057  the restrictions on partial matching have been lifted. It remains for backwards  32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the
1058    restrictions on partial matching have been lifted. It remains for backwards
1059  compatibility. */  compatibility. */
1060    
1061  #define PCRE_MODE8         0x0001  /* compiled in 8 bit mode */  #define PCRE_MODE8         0x00000001  /* compiled in 8 bit mode */
1062  #define PCRE_MODE16        0x0002  /* compiled in 16 bit mode */  #define PCRE_MODE16        0x00000002  /* compiled in 16 bit mode */
1063  #define PCRE_MODE32        0x0004  /* compiled in 32 bit mode */  #define PCRE_MODE32        0x00000004  /* compiled in 32 bit mode */
1064  #define PCRE_FIRSTSET      0x0010  /* first_char is set */  #define PCRE_FIRSTSET      0x00000010  /* first_char is set */
1065  #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */  #define PCRE_FCH_CASELESS  0x00000020  /* caseless first char */
1066  #define PCRE_REQCHSET      0x0040  /* req_byte is set */  #define PCRE_REQCHSET      0x00000040  /* req_byte is set */
1067  #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */  #define PCRE_RCH_CASELESS  0x00000080  /* caseless requested char */
1068  #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */  #define PCRE_STARTLINE     0x00000100  /* start after \n for multiline */
1069  #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x00000200  /* can't use partial with this regex */
1070  #define PCRE_JCHANGED      0x0400  /* j option used in regex */  #define PCRE_JCHANGED      0x00000400  /* j option used in regex */
1071  #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x00000800  /* explicit \r or \n in pattern */
1072  #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */  #define PCRE_HASTHEN       0x00001000  /* pattern contains (*THEN) */
1073    #define PCRE_MLSET         0x00002000  /* match limit set by regex */
1074    #define PCRE_RLSET         0x00004000  /* recursion limit set by regex */
1075    #define PCRE_MATCH_EMPTY   0x00008000  /* pattern can match empty string */
1076    
1077  #if defined COMPILE_PCRE8  #if defined COMPILE_PCRE8
1078  #define PCRE_MODE          PCRE_MODE8  #define PCRE_MODE          PCRE_MODE8
# Line 1153  time, run time, or study time, respectiv Line 1097  time, run time, or study time, respectiv
1097  #define PUBLIC_COMPILE_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
1098    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
1099     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1100     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESS| \
1101       PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1102     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1103     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1104    
1105  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
1106    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
# Line 1172  time, run time, or study time, respectiv Line 1117  time, run time, or study time, respectiv
1117     (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \     (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
1118      PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)      PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
1119    
1120    #define PUBLIC_JIT_EXEC_OPTIONS \
1121       (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\
1122        PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD)
1123    
1124  /* Magic number to provide a small check against being handed junk. */  /* Magic number to provide a small check against being handed junk. */
1125    
1126  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
# Line 1275  a positive value. */ Line 1224  a positive value. */
1224    
1225  /* The remaining definitions work in both environments. */  /* The remaining definitions work in both environments. */
1226    
1227    #define CHAR_NULL                   '\0'
1228  #define CHAR_HT                     '\t'  #define CHAR_HT                     '\t'
1229  #define CHAR_VT                     '\v'  #define CHAR_VT                     '\v'
1230  #define CHAR_FF                     '\f'  #define CHAR_FF                     '\f'
# Line 1506  a positive value. */ Line 1456  a positive value. */
1456  #define STRING_xdigit               "xdigit"  #define STRING_xdigit               "xdigit"
1457    
1458  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
1459    #define STRING_WEIRD_STARTWORD      "[:<:]]"
1460    #define STRING_WEIRD_ENDWORD        "[:>:]]"
1461    
1462  #define STRING_CR_RIGHTPAR             "CR)"  #define STRING_CR_RIGHTPAR              "CR)"
1463  #define STRING_LF_RIGHTPAR             "LF)"  #define STRING_LF_RIGHTPAR              "LF)"
1464  #define STRING_CRLF_RIGHTPAR           "CRLF)"  #define STRING_CRLF_RIGHTPAR            "CRLF)"
1465  #define STRING_ANY_RIGHTPAR            "ANY)"  #define STRING_ANY_RIGHTPAR             "ANY)"
1466  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR         "ANYCRLF)"
1467  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR     "BSR_ANYCRLF)"
1468  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR     "BSR_UNICODE)"
1469  #ifdef COMPILE_PCRE8  #define STRING_UTF8_RIGHTPAR            "UTF8)"
1470  #define STRING_UTF_RIGHTPAR            "UTF8)"  #define STRING_UTF16_RIGHTPAR           "UTF16)"
1471  #endif  #define STRING_UTF32_RIGHTPAR           "UTF32)"
1472  #ifdef COMPILE_PCRE16  #define STRING_UTF_RIGHTPAR             "UTF)"
1473  #define STRING_UTF_RIGHTPAR            "UTF16)"  #define STRING_UCP_RIGHTPAR             "UCP)"
1474  #endif  #define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)"
1475  #ifdef COMPILE_PCRE32  #define STRING_NO_START_OPT_RIGHTPAR    "NO_START_OPT)"
1476  #define STRING_UTF_RIGHTPAR            "UTF32)"  #define STRING_LIMIT_MATCH_EQ           "LIMIT_MATCH="
1477  #endif  #define STRING_LIMIT_RECURSION_EQ       "LIMIT_RECURSION="
 #define STRING_UCP_RIGHTPAR            "UCP)"  
 #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"  
1478    
1479  #else  /* SUPPORT_UTF */  #else  /* SUPPORT_UTF */
1480    
# Line 1544  only. */ Line 1494  only. */
1494  #define CHAR_ESC                    '\033'  #define CHAR_ESC                    '\033'
1495  #define CHAR_DEL                    '\177'  #define CHAR_DEL                    '\177'
1496    
1497    #define CHAR_NULL                   '\0'
1498  #define CHAR_SPACE                  '\040'  #define CHAR_SPACE                  '\040'
1499  #define CHAR_EXCLAMATION_MARK       '\041'  #define CHAR_EXCLAMATION_MARK       '\041'
1500  #define CHAR_QUOTATION_MARK         '\042'  #define CHAR_QUOTATION_MARK         '\042'
# Line 1771  only. */ Line 1722  only. */
1722  #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t  #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1723    
1724  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1725    #define STRING_WEIRD_STARTWORD      STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1726    #define STRING_WEIRD_ENDWORD        STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1727    
1728  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR              STR_C STR_R STR_RIGHT_PARENTHESIS
1729  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR              STR_L STR_F STR_RIGHT_PARENTHESIS
1730  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR            STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1731  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR             STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1732  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR         STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1733  #define STRING_BSR_ANYCRLF_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_BSR_ANYCRLF_RIGHTPAR     STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1734  #define STRING_BSR_UNICODE_RIGHTPAR    STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS  #define STRING_BSR_UNICODE_RIGHTPAR     STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1735  #ifdef COMPILE_PCRE8  #define STRING_UTF8_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1736  #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF16_RIGHTPAR           STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1737  #endif  #define STRING_UTF32_RIGHTPAR           STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1738  #ifdef COMPILE_PCRE16  #define STRING_UTF_RIGHTPAR             STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1739  #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS  #define STRING_UCP_RIGHTPAR             STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1740  #endif  #define STRING_NO_AUTO_POSSESS_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_A STR_U STR_T STR_O STR_UNDERSCORE STR_P STR_O STR_S STR_S STR_E STR_S STR_S STR_RIGHT_PARENTHESIS
1741  #ifdef COMPILE_PCRE32  #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
1742  #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS  #define STRING_LIMIT_MATCH_EQ           STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN
1743  #endif  #define STRING_LIMIT_RECURSION_EQ       STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN
 #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS  
 #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  
1744    
1745  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
1746    
# Line 1830  only. */ Line 1781  only. */
1781  #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */  #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1782  #define PT_WORD       8    /* Word - L plus N plus underscore */  #define PT_WORD       8    /* Word - L plus N plus underscore */
1783  #define PT_CLIST      9    /* Pseudo-property: match character list */  #define PT_CLIST      9    /* Pseudo-property: match character list */
1784    #define PT_UCNC      10    /* Universal Character nameable character */
1785    #define PT_TABSIZE   11    /* Size of square table for autopossessify tests */
1786    
1787    /* The following special properties are used only in XCLASS items, when POSIX
1788    classes are specified and PCRE_UCP is set - in other words, for Unicode
1789    handling of these classes. They are not available via the \p or \P escapes like
1790    those in the above list, and so they do not take part in the autopossessifying
1791    table. */
1792    
1793    #define PT_PXGRAPH   11    /* [:graph:] - characters that mark the paper */
1794    #define PT_PXPRINT   12    /* [:print:] - [:graph:] plus non-control spaces */
1795    #define PT_PXPUNCT   13    /* [:punct:] - punctuation characters */
1796    
1797  /* 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
1798  contain characters with values greater than 255. */  contain characters with values greater than 255. */
1799    
1800  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT       0x01    /* Flag: this is a negative class */
1801  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP       0x02    /* Flag: a 32-byte map is present */
1802    #define XCL_HASPROP   0x04    /* Flag: property checks are present. */
1803    
1804  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
1805  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
# Line 1844  contain characters with values greater t Line 1808  contain characters with values greater t
1808  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1809    
1810  /* 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
1811  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 0
1812  0 for a data character.  Also, they must appear in the same order as in the opcode  for a data character.  Also, they must appear in the same order as in the
1813  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it  opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1814  corresponds to "." in DOTALL mode rather than an escape sequence. It is also  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1815  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1816  non-DOTALL mode, "." behaves like \N.  non-DOTALL mode, "." behaves like \N.
# Line 1869  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E Line 1833  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, E
1833         ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_E, ESC_Q, ESC_g, ESC_k,
1834         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };         ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1835    
 /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to  
 OP_EOD must correspond in order to the list of escapes immediately above.  
1836    
1837  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  /********************** Opcode definitions ******************/
1838  that follow must also be updated to match. There are also tables called  
1839  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  /****** NOTE NOTE NOTE ******
1840    
1841    Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1842    order to the list of escapes immediately above. Furthermore, values up to
1843    OP_DOLLM must not be changed without adjusting the table called autoposstab in
1844    pcre_compile.c
1845    
1846    Whenever this list is updated, the two macro definitions that follow must be
1847    updated to match. The possessification table called "opcode_possessify" in
1848    pcre_compile.c must also be updated, and also the tables called "coptable"
1849    and "poptable" in pcre_dfa_exec.c.
1850    
1851    ****** NOTE NOTE NOTE ******/
1852    
1853    
1854    /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1855    are used in a table for deciding whether a repeated character type can be
1856    auto-possessified. */
1857    
1858    #define FIRST_AUTOTAB_OP       OP_NOT_DIGIT
1859    #define LAST_AUTOTAB_LEFT_OP   OP_EXTUNI
1860    #define LAST_AUTOTAB_RIGHT_OP  OP_DOLLM
1861    
1862  enum {  enum {
1863    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 1907  enum { Line 1890  enum {
1890    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */    OP_EODN,           /* 23 End of data or \n at end of data (\Z) */
1891    OP_EOD,            /* 24 End of data (\z) */    OP_EOD,            /* 24 End of data (\z) */
1892    
1893    OP_CIRC,           /* 25 Start of line - not multiline */    /* Line end assertions */
1894    OP_CIRCM,          /* 26 Start of line - multiline */  
1895    OP_DOLL,           /* 27 End of line - not multiline */    OP_DOLL,           /* 25 End of line - not multiline */
1896    OP_DOLLM,          /* 28 End of line - multiline */    OP_DOLLM,          /* 26 End of line - multiline */
1897      OP_CIRC,           /* 27 Start of line - not multiline */
1898      OP_CIRCM,          /* 28 Start of line - multiline */
1899    
1900      /* Single characters; caseful must precede the caseless ones */
1901    
1902    OP_CHAR,           /* 29 Match one character, casefully */    OP_CHAR,           /* 29 Match one character, casefully */
1903    OP_CHARI,          /* 30 Match one character, caselessly */    OP_CHARI,          /* 30 Match one character, caselessly */
1904    OP_NOT,            /* 31 Match one character, not the given one, casefully */    OP_NOT,            /* 31 Match one character, not the given one, casefully */
# Line 1919  enum { Line 1907  enum {
1907    /* The following sets of 13 opcodes must always be kept in step because    /* The following sets of 13 opcodes must always be kept in step because
1908    the offset from the first one is used to generate the others. */    the offset from the first one is used to generate the others. */
1909    
1910    /**** Single characters, caseful, must precede the caseless ones ****/    /* Repeated characters; caseful must precede the caseless ones */
1911    
1912    OP_STAR,           /* 33 The maximizing and minimizing versions of */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1913    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
# Line 1937  enum { Line 1925  enum {
1925    OP_POSQUERY,       /* 44 Posesssified query, caseful */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1926    OP_POSUPTO,        /* 45 Possessified upto, caseful */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1927    
1928    /**** Single characters, caseless, must follow the caseful ones */    /* Repeated characters; caseless must follow the caseful ones */
1929    
1930    OP_STARI,          /* 46 */    OP_STARI,          /* 46 */
1931    OP_MINSTARI,       /* 47 */    OP_MINSTARI,       /* 47 */
# Line 1955  enum { Line 1943  enum {
1943    OP_POSQUERYI,      /* 57 Posesssified query, caseless */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1944    OP_POSUPTOI,       /* 58 Possessified upto, caseless */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1945    
1946    /**** The negated ones must follow the non-negated ones, and match them ****/    /* The negated ones must follow the non-negated ones, and match them */
1947    /**** Negated single character, caseful; must precede the caseless ones ****/    /* Negated repeated character, caseful; must precede the caseless ones */
1948    
1949    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1950    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
# Line 1974  enum { Line 1962  enum {
1962    OP_NOTPOSQUERY,    /* 70 */    OP_NOTPOSQUERY,    /* 70 */
1963    OP_NOTPOSUPTO,     /* 71 */    OP_NOTPOSUPTO,     /* 71 */
1964    
1965    /**** Negated single character, caseless; must follow the caseful ones ****/    /* Negated repeated character, caseless; must follow the caseful ones */
1966    
1967    OP_NOTSTARI,       /* 72 */    OP_NOTSTARI,       /* 72 */
1968    OP_NOTMINSTARI,    /* 73 */    OP_NOTMINSTARI,    /* 73 */
# Line 1992  enum { Line 1980  enum {
1980    OP_NOTPOSQUERYI,   /* 83 */    OP_NOTPOSQUERYI,   /* 83 */
1981    OP_NOTPOSUPTOI,    /* 84 */    OP_NOTPOSUPTOI,    /* 84 */
1982    
1983    /**** Character types ****/    /* Character types */
1984    
1985    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1986    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
# Line 2023  enum { Line 2011  enum {
2011    OP_CRRANGE,        /* 104 These are different to the three sets above. */    OP_CRRANGE,        /* 104 These are different to the three sets above. */
2012    OP_CRMINRANGE,     /* 105 */    OP_CRMINRANGE,     /* 105 */
2013    
2014      OP_CRPOSSTAR,      /* 106 Possessified versions */
2015      OP_CRPOSPLUS,      /* 107 */
2016      OP_CRPOSQUERY,     /* 108 */
2017      OP_CRPOSRANGE,     /* 109 */
2018    
2019    /* End of quantifier opcodes */    /* End of quantifier opcodes */
2020    
2021    OP_CLASS,          /* 106 Match a character class, chars < 256 only */    OP_CLASS,          /* 110 Match a character class, chars < 256 only */
2022    OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 111 Same, but the bitmap was created from a negative
2023                                class - the difference is relevant only when a                                class - the difference is relevant only when a
2024                                character > 255 is encountered. */                                character > 255 is encountered. */
2025    OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the    OP_XCLASS,         /* 112 Extended class for handling > 255 chars within the
2026                                class. This does both positive and negative. */                                class. This does both positive and negative. */
2027    OP_REF,            /* 109 Match a back reference, casefully */    OP_REF,            /* 113 Match a back reference, casefully */
2028    OP_REFI,           /* 110 Match a back reference, caselessly */    OP_REFI,           /* 114 Match a back reference, caselessly */
2029    OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */    OP_DNREF,          /* 115 Match a duplicate name backref, casefully */
2030    OP_CALLOUT,        /* 112 Call out to external function if provided */    OP_DNREFI,         /* 116 Match a duplicate name backref, caselessly */
2031      OP_RECURSE,        /* 117 Match a numbered subpattern (possibly recursive) */
2032    OP_ALT,            /* 113 Start of alternation */    OP_CALLOUT,        /* 118 Call out to external function if provided */
2033    OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */  
2034    OP_KETRMAX,        /* 115 These two must remain together and in this */    OP_ALT,            /* 119 Start of alternation */
2035    OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */    OP_KET,            /* 120 End of group that doesn't have an unbounded repeat */
2036    OP_KETRPOS,        /* 117 Possessive unlimited repeat. */    OP_KETRMAX,        /* 121 These two must remain together and in this */
2037      OP_KETRMIN,        /* 122 order. They are for groups the repeat for ever. */
2038      OP_KETRPOS,        /* 123 Possessive unlimited repeat. */
2039    
2040    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four    /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
2041    asserts must remain in order. */    asserts must remain in order. */
2042    
2043    OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 124 Move pointer back - used in lookbehind assertions */
2044    OP_ASSERT,         /* 119 Positive lookahead */    OP_ASSERT,         /* 125 Positive lookahead */
2045    OP_ASSERT_NOT,     /* 120 Negative lookahead */    OP_ASSERT_NOT,     /* 126 Negative lookahead */
2046    OP_ASSERTBACK,     /* 121 Positive lookbehind */    OP_ASSERTBACK,     /* 127 Positive lookbehind */
2047    OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2048    
2049    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately    /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2050    after the assertions, with ONCE first, as there's a test for >= ONCE for a    after the assertions, with ONCE first, as there's a test for >= ONCE for a
2051    subpattern that isn't an assertion. The POS versions must immediately follow    subpattern that isn't an assertion. The POS versions must immediately follow
2052    the non-POS versions in each case. */    the non-POS versions in each case. */
2053    
2054    OP_ONCE,           /* 123 Atomic group, contains captures */    OP_ONCE,           /* 129 Atomic group, contains captures */
2055    OP_ONCE_NC,        /* 124 Atomic group containing no captures */    OP_ONCE_NC,        /* 130 Atomic group containing no captures */
2056    OP_BRA,            /* 125 Start of non-capturing bracket */    OP_BRA,            /* 131 Start of non-capturing bracket */
2057    OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */    OP_BRAPOS,         /* 132 Ditto, with unlimited, possessive repeat */
2058    OP_CBRA,           /* 127 Start of capturing bracket */    OP_CBRA,           /* 133 Start of capturing bracket */
2059    OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */    OP_CBRAPOS,        /* 134 Ditto, with unlimited, possessive repeat */
2060    OP_COND,           /* 129 Conditional group */    OP_COND,           /* 135 Conditional group */
2061    
2062    /* These five must follow the previous five, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
2063    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
2064    
2065    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 136 Start of non-capturing bracket, check empty  */
2066    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */    OP_SBRAPOS,        /* 137 Ditto, with unlimited, possessive repeat */
2067    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */    OP_SCBRA,          /* 138 Start of capturing bracket, check empty */
2068    OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */    OP_SCBRAPOS,       /* 139 Ditto, with unlimited, possessive repeat */
2069    OP_SCOND,          /* 134 Conditional group, check empty */    OP_SCOND,          /* 140 Conditional group, check empty */
2070    
2071    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
2072    
2073    OP_CREF,           /* 135 Used to hold a capture number as condition */    OP_CREF,           /* 141 Used to hold a capture number as condition */
2074    OP_NCREF,          /* 136 Same, but generated by a name reference*/    OP_DNCREF,         /* 142 Used to point to duplicate names as a condition */
2075    OP_RREF,           /* 137 Used to hold a recursion number as condition */    OP_RREF,           /* 143 Used to hold a recursion number as condition */
2076    OP_NRREF,          /* 138 Same, but generated by a name reference*/    OP_DNRREF,         /* 144 Used to point to duplicate names as a condition */
2077    OP_DEF,            /* 139 The DEFINE condition */    OP_DEF,            /* 145 The DEFINE condition */
2078    
2079    OP_BRAZERO,        /* 140 These two must remain together and in this */    OP_BRAZERO,        /* 146 These two must remain together and in this */
2080    OP_BRAMINZERO,     /* 141 order. */    OP_BRAMINZERO,     /* 147 order. */
2081    OP_BRAPOSZERO,     /* 142 */    OP_BRAPOSZERO,     /* 148 */
2082    
2083    /* These are backtracking control verbs */    /* These are backtracking control verbs */
2084    
2085    OP_MARK,           /* 143 always has an argument */    OP_MARK,           /* 149 always has an argument */
2086    OP_PRUNE,          /* 144 */    OP_PRUNE,          /* 150 */
2087    OP_PRUNE_ARG,      /* 145 same, but with argument */    OP_PRUNE_ARG,      /* 151 same, but with argument */
2088    OP_SKIP,           /* 146 */    OP_SKIP,           /* 152 */
2089    OP_SKIP_ARG,       /* 147 same, but with argument */    OP_SKIP_ARG,       /* 153 same, but with argument */
2090    OP_THEN,           /* 148 */    OP_THEN,           /* 154 */
2091    OP_THEN_ARG,       /* 149 same, but with argument */    OP_THEN_ARG,       /* 155 same, but with argument */
2092    OP_COMMIT,         /* 150 */    OP_COMMIT,         /* 156 */
2093    
2094    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
2095    
2096    OP_FAIL,           /* 151 */    OP_FAIL,           /* 157 */
2097    OP_ACCEPT,         /* 152 */    OP_ACCEPT,         /* 158 */
2098    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */    OP_ASSERT_ACCEPT,  /* 159 Used inside assertions */
2099    OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */    OP_CLOSE,          /* 160 Used before OP_ACCEPT to close open captures */
2100    
2101    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
2102    
2103    OP_SKIPZERO,       /* 155 */    OP_SKIPZERO,       /* 161 */
2104    
2105    /* This is not an opcode, but is used to check that tables indexed by opcode    /* This is not an opcode, but is used to check that tables indexed by opcode
2106    are the correct length, in order to catch updating errors - there have been    are the correct length, in order to catch updating errors - there have been
# Line 2116  enum { Line 2111  enum {
2111    
2112  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2113  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
2114  called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */  called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in
2115    pcre_dfa_exec.c that must be updated. */
2116    
2117    
2118  /* 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
# Line 2129  some cases doesn't actually use these na Line 2125  some cases doesn't actually use these na
2125    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
2126    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
2127    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
2128    "^", "^", "$", "$", "char", "chari", "not", "noti",             \    "$", "$", "^", "^", "char", "chari", "not", "noti",             \
2129    "*", "*?", "+", "+?", "?", "??",                                \    "*", "*?", "+", "+?", "?", "??",                                \
2130    "{", "{", "{",                                                  \    "{", "{", "{",                                                  \
2131    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 2145  some cases doesn't actually use these na Line 2141  some cases doesn't actually use these na
2141    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
2142    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
2143    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
2144    "class", "nclass", "xclass", "Ref", "Refi",                     \    "*+","++", "?+", "{",                                           \
2145      "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi",  \
2146    "Recurse", "Callout",                                           \    "Recurse", "Callout",                                           \
2147    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
2148    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
# Line 2154  some cases doesn't actually use these na Line 2151  some cases doesn't actually use these na
2151    "Cond",                                                         \    "Cond",                                                         \
2152    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \    "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
2153    "SCond",                                                        \    "SCond",                                                        \
2154    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2155    "Brazero", "Braminzero", "Braposzero",                          \    "Brazero", "Braminzero", "Braposzero",                          \
2156    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
2157    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \    "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
# Line 2179  in UTF-8 mode. The code that uses this t Line 2176  in UTF-8 mode. The code that uses this t
2176    3, 3,                          /* \P, \p                                 */ \    3, 3,                          /* \P, \p                                 */ \
2177    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
2178    1,                             /* \X                                     */ \    1,                             /* \X                                     */ \
2179    1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \    1, 1, 1, 1, 1, 1,              /* \Z, \z, $, $M ^, ^M                    */ \
2180    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
2181    2,                             /* Chari  - the minimum length            */ \    2,                             /* Chari  - the minimum length            */ \
2182    2,                             /* not                                    */ \    2,                             /* not                                    */ \
# Line 2210  in UTF-8 mode. The code that uses this t Line 2207  in UTF-8 mode. The code that uses this t
2207    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
2208    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
2209    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
2210      1, 1, 1, 1+2*IMM2_SIZE,        /* Possessive *+, ++, ?+, CRPOSRANGE      */ \
2211    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
2212    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
2213    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
2214    1+IMM2_SIZE,                   /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
2215    1+IMM2_SIZE,                   /* REFI                                   */ \    1+IMM2_SIZE,                   /* REFI                                   */ \
2216      1+2*IMM2_SIZE,                 /* DNREF                                  */ \
2217      1+2*IMM2_SIZE,                 /* DNREFI                                 */ \
2218    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
2219    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
2220    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
# Line 2239  in UTF-8 mode. The code that uses this t Line 2239  in UTF-8 mode. The code that uses this t
2239    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
2240    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
2241    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
2242    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* CREF, DNCREF                           */ \
2243    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \    1+IMM2_SIZE, 1+2*IMM2_SIZE,    /* RREF, DNRREF                           */ \
2244    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
2245    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
2246    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
# Line 2249  in UTF-8 mode. The code that uses this t Line 2249  in UTF-8 mode. The code that uses this t
2249    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \    1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
2250    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */    1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
2251    
2252  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
 condition. */  
2253    
2254  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
2255    
# Line 2265  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 2264  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
2264         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2265         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2266         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2267         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERRCOUNT };         ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2268           ERR80, ERR81, ERR82, ERR83, ERR84, ERRCOUNT };
2269    
2270  /* JIT compiling modes. The function list is indexed by them. */  /* JIT compiling modes. The function list is indexed by them. */
2271    
2272  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2273         JIT_NUMBER_OF_COMPILE_MODES };         JIT_NUMBER_OF_COMPILE_MODES };
2274    
# Line 2275  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COM Line 2276  enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COM
2276  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
2277  offset to the name table so that if a regex is compiled on one host, saved, and  offset to the name table so that if a regex is compiled on one host, saved, and
2278  then run on another where the size of pointers is different, all might still  then run on another where the size of pointers is different, all might still
2279  be well. For the case of compiled-on-4 and run-on-8, we include an extra  be well.
 pointer that is always NULL. For future-proofing, a few dummy fields were  
 originally included - even though you can never get this planning right - but  
 there is only one left now.  
   
 NOTE NOTE NOTE:  
 Because people can now save and re-use compiled patterns, any additions to this  
 structure should be made at the end, and something earlier (e.g. a new  
 flag in the options or one of the dummy fields) should indicate that the new  
 fields are present. Currently PCRE always sets the dummy fields to zero.  
 NOTE NOTE NOTE  
 */  
2280    
2281  #if defined COMPILE_PCRE8  The size of the structure must be a multiple of 8 bytes. For the case of
2282  #define REAL_PCRE real_pcre  compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so
2283  #elif defined COMPILE_PCRE16  that there are an even number of pointers which therefore are a multiple of 8
2284  #define REAL_PCRE real_pcre16  bytes.
2285  #elif defined COMPILE_PCRE32  
2286  #define REAL_PCRE real_pcre32  It is necessary to fork the struct for the 32 bit library, since it needs to
2287  #endif  use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the
2288    typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit
2289  /* It is necessary to fork the struct for 32 bit, since it needs to use  variants.
2290   * pcre_uchar for first_char and req_char. Can't put an ifdef inside the  
2291   * typedef since pcretest needs access to  the struct of the 8-, 16-  *** WARNING ***
2292   * and 32-bit variants. */  When new fields are added to these structures, remember to adjust the code in
2293    pcre_byte_order.c that is concerned with swapping the byte order of the fields
2294    when a compiled regex is reloaded on a host with different endianness.
2295    *** WARNING ***
2296    There is also similar byte-flipping code in pcretest.c, which is used for
2297    testing the byte-flipping features. It must also be kept in step.
2298    *** WARNING ***
2299    */
2300    
2301  typedef struct real_pcre8_or_16 {  typedef struct real_pcre8_or_16 {
2302    pcre_uint32 magic_number;    pcre_uint32 magic_number;
2303    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2304    pcre_uint32 options;            /* Public options */    pcre_uint32 options;            /* Public options */
2305    pcre_uint16 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
2306      pcre_uint32 limit_match;        /* Limit set from regex */
2307      pcre_uint32 limit_recursion;    /* Limit set from regex */
2308      pcre_uint16 first_char;         /* Starting character */
2309      pcre_uint16 req_char;           /* This character must be seen */
2310    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2311    pcre_uint16 top_bracket;        /* Highest numbered group */    pcre_uint16 top_bracket;        /* Highest numbered group */
2312    pcre_uint16 top_backref;        /* Highest numbered back reference */    pcre_uint16 top_backref;        /* Highest numbered back reference */
   pcre_uint16 first_char;         /* Starting character */  
   pcre_uint16 req_char;           /* This character must be seen */  
2313    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2314    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
2315    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
2316    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
2317      pcre_uint16 dummy1;             /* To ensure size is a multiple of 8 */
2318      pcre_uint16 dummy2;             /* To ensure size is a multiple of 8 */
2319      pcre_uint16 dummy3;             /* To ensure size is a multiple of 8 */
2320    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2321    const pcre_uint8 *nullpad;      /* NULL padding */    void             *nullpad;      /* NULL padding */
2322  } real_pcre8_or_16;  } real_pcre8_or_16;
2323    
2324  typedef struct real_pcre8_or_16 real_pcre;  typedef struct real_pcre8_or_16 real_pcre;
# Line 2326  typedef struct real_pcre32 { Line 2328  typedef struct real_pcre32 {
2328    pcre_uint32 magic_number;    pcre_uint32 magic_number;
2329    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2330    pcre_uint32 options;            /* Public options */    pcre_uint32 options;            /* Public options */
2331    pcre_uint16 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
2332      pcre_uint32 limit_match;        /* Limit set from regex */
2333      pcre_uint32 limit_recursion;    /* Limit set from regex */
2334      pcre_uint32 first_char;         /* Starting character */
2335      pcre_uint32 req_char;           /* This character must be seen */
2336    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2337    pcre_uint16 top_bracket;        /* Highest numbered group */    pcre_uint16 top_bracket;        /* Highest numbered group */
2338    pcre_uint16 top_backref;        /* Highest numbered back reference */    pcre_uint16 top_backref;        /* Highest numbered back reference */
   pcre_uint32 first_char;         /* Starting character */  
   pcre_uint32 req_char;           /* This character must be seen */  
2339    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2340    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
2341    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
2342    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
2343    pcre_uint16 dummy1;             /* for later expansion */    pcre_uint16 dummy;              /* To ensure size is a multiple of 8 */
   pcre_uint16 dummy2;             /* for later expansion */  
2344    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2345    void *nullpad;                  /* for later expansion */    void             *nullpad;      /* NULL padding */
2346  } real_pcre32;  } real_pcre32;
2347    
2348    #if defined COMPILE_PCRE8
2349    #define REAL_PCRE real_pcre
2350    #elif defined COMPILE_PCRE16
2351    #define REAL_PCRE real_pcre16
2352    #elif defined COMPILE_PCRE32
2353    #define REAL_PCRE real_pcre32
2354    #endif
2355    
2356  /* Assert that the size of REAL_PCRE is divisible by 8 */  /* Assert that the size of REAL_PCRE is divisible by 8 */
2357  typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];  typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
2358    
# Line 2375  typedef struct open_capitem { Line 2386  typedef struct open_capitem {
2386    pcre_uint16 flag;             /* Set TRUE if recursive back ref */    pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2387  } open_capitem;  } open_capitem;
2388    
2389    /* Structure for building a list of named groups during the first pass of
2390    compiling. */
2391    
2392    typedef struct named_group {
2393      const pcre_uchar  *name;          /* Points to the name in the pattern */
2394      int                length;        /* Length of the name */
2395      pcre_uint32        number;        /* Group number */
2396    } named_group;
2397    
2398  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2399  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2400    
# Line 2387  typedef struct compile_data { Line 2407  typedef struct compile_data {
2407    const pcre_uchar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2408    const pcre_uchar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2409    const pcre_uchar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
   open_capitem *open_caps;          /* Chain of open capture items */  
2410    pcre_uchar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
2411      open_capitem *open_caps;          /* Chain of open capture items */
2412      named_group *named_groups;        /* Points to vector in pre-compile */
2413    pcre_uchar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
2414    int  names_found;                 /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
2415    int  name_entry_size;             /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
2416      int  named_group_list_size;       /* Number of entries in the list */
2417    int  workspace_size;              /* Size of workspace */    int  workspace_size;              /* Size of workspace */
2418    int  bracount;                    /* Count of capturing parens as we compile */    unsigned int bracount;            /* Count of capturing parens as we compile */
2419    int  final_bracount;              /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
2420    int  max_lookbehind;              /* Maximum lookbehind (characters) */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2421    int  top_backref;                 /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
2422    unsigned int backref_map;         /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
2423      unsigned int namedrefcount;       /* Number of backreferences by name */
2424      int  parens_depth;                /* Depth of nested parentheses */
2425    int  assert_depth;                /* Depth of nested assertions */    int  assert_depth;                /* Depth of nested assertions */
2426    int  external_options;            /* External (initial) options */    pcre_uint32 external_options;     /* External (initial) options */
2427    int  external_flags;              /* External flag bits to be set */    pcre_uint32 external_flags;       /* External flag bits to be set */
2428    int  req_varyopt;                 /* "After variable item" flag for reqbyte */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2429    BOOL had_accept;                  /* (*ACCEPT) encountered */    BOOL had_accept;                  /* (*ACCEPT) encountered */
2430    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */    BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2431    BOOL check_lookbehind;            /* Lookbehinds need later checking */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
2432      BOOL dupnames;                    /* Duplicate names exist */
2433    int  nltype;                      /* Newline type */    int  nltype;                      /* Newline type */
2434    int  nllen;                       /* Newline string length */    int  nllen;                       /* Newline string length */
2435    pcre_uchar nl[4];                 /* Newline string when fixed length */    pcre_uchar nl[4];                 /* Newline string when fixed length */
# Line 2423  call within the pattern; used by pcre_ex Line 2448  call within the pattern; used by pcre_ex
2448    
2449  typedef struct recursion_info {  typedef struct recursion_info {
2450    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2451    int group_num;                  /* Number of group that was called */    unsigned int group_num;         /* Number of group that was called */
2452    int *offset_save;               /* Pointer to start of saved offsets */    int *offset_save;               /* Pointer to start of saved offsets */
2453    int saved_max;                  /* Number of saved offsets */    int saved_max;                  /* Number of saved offsets */
2454      int saved_capture_last;         /* Last capture number */
2455    PCRE_PUCHAR subject_position;   /* Position at start of recursion */    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
2456  } recursion_info;  } recursion_info;
2457    
# Line 2462  typedef struct match_data { Line 2488  typedef struct match_data {
2488    int    nllen;                   /* Newline string length */    int    nllen;                   /* Newline string length */
2489    int    name_count;              /* Number of names in name table */    int    name_count;              /* Number of names in name table */
2490    int    name_entry_size;         /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
2491      unsigned int skip_arg_count;    /* For counting SKIP_ARGs */
2492      unsigned int ignore_skip_arg;   /* For re-run when SKIP arg name not found */
2493    pcre_uchar *name_table;         /* Table of names */    pcre_uchar *name_table;         /* Table of names */
2494    pcre_uchar nl[4];               /* Newline string when fixed */    pcre_uchar nl[4];               /* Newline string when fixed */
2495    const  pcre_uint8 *lcc;         /* Points to lower casing table */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2496    const  pcre_uint8 *fcc;         /* Points to case-flipping table */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2497    const  pcre_uint8 *ctypes;      /* Points to table of type maps */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
   BOOL   offset_overflow;         /* Set if too many extractions */  
2498    BOOL   notbol;                  /* NOTBOL flag */    BOOL   notbol;                  /* NOTBOL flag */
2499    BOOL   noteol;                  /* NOTEOL flag */    BOOL   noteol;                  /* NOTEOL flag */
2500    BOOL   utf;                     /* UTF-8 / UTF-16 flag */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
# Line 2479  typedef struct match_data { Line 2506  typedef struct match_data {
2506    BOOL   hitend;                  /* Hit the end of the subject at some point */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2507    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2508    BOOL   hasthen;                 /* Pattern contains (*THEN) */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
   BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */  
2509    const  pcre_uchar *start_code;  /* For use when recursing */    const  pcre_uchar *start_code;  /* For use when recursing */
2510    PCRE_PUCHAR start_subject;      /* Start of the subject string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2511    PCRE_PUCHAR end_subject;        /* End of the subject string */    PCRE_PUCHAR end_subject;        /* End of the subject string */
# Line 2488  typedef struct match_data { Line 2514  typedef struct match_data {
2514    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2515    int    partial;                 /* PARTIAL options */    int    partial;                 /* PARTIAL options */
2516    int    end_offset_top;          /* Highwater mark at end of match */    int    end_offset_top;          /* Highwater mark at end of match */
2517    int    capture_last;            /* Most recent capture number */    pcre_int32 capture_last;        /* Most recent capture number + overflow flag */
2518    int    start_offset;            /* The start offset value */    int    start_offset;            /* The start offset value */
2519    int    match_function_type;     /* Set for certain special calls of MATCH() */    int    match_function_type;     /* Set for certain special calls of MATCH() */
2520    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
# Line 2681  extern int               PRIV(strcmp_uc_ Line 2707  extern int               PRIV(strcmp_uc_
2707  extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);  extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2708  extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,  extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2709                             int *, BOOL);                             int *, BOOL);
2710  extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);  extern unsigned int      PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2711  extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);  extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2712  extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,  extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2713                             int *, BOOL);                             int *, BOOL);
# Line 2690  extern BOOL              PRIV(xclass)(pc Line 2716  extern BOOL              PRIV(xclass)(pc
2716  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
2717  extern void              PRIV(jit_compile)(const REAL_PCRE *,  extern void              PRIV(jit_compile)(const REAL_PCRE *,
2718                             PUBL(extra) *, int);                             PUBL(extra) *, int);
2719  extern int               PRIV(jit_exec)(const REAL_PCRE *, const PUBL(extra) *,  extern int               PRIV(jit_exec)(const PUBL(extra) *,
2720                             const pcre_uchar *, int, int, int, int *, int);                             const pcre_uchar *, int, int, int, int *, int);
2721  extern void              PRIV(jit_free)(void *);  extern void              PRIV(jit_free)(void *);
2722  extern int               PRIV(jit_get_size)(void *);  extern int               PRIV(jit_get_size)(void *);

Legend:
Removed from v.1128  
changed lines
  Added in v.1425

  ViewVC Help
Powered by ViewVC 1.1.5