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

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

Parent Directory | Revision Log | View Patch Patch

-code/trunk/pcre_internal.h
revision 393 by ph10,
Tue Mar 17 21:47:54 2009 UTC
+code/branches/pcre16/pcre_internal.h
revision 769 by zherczeg,
Sun Nov 27 17:40:33 2011 UTC
 Line 7
  and semantics are as close as possible to those of the Perl 5 language.
                         Written by Philip Hazel
-            Copyright (c) 1997-2009 University of Cambridge
+            Copyright (c) 1997-2011 University of Cambridge
  -----------------------------------------------------------------------------
  Redistribution and use in source and binary forms, with or without
 Line 40 
 POSSIBILITY OF SUCH DAMAGE.
  /* This header contains definitions that are shared between the different
  modules, but which are not relevant to the exported API. This includes some
- functions whose names all begin with "_pcre_". */
+ functions whose names all begin with "_pcre_" or "_pcre16_" depending on
+ the PRIV macro. */
  #ifndef PCRE_INTERNAL_H
  #define PCRE_INTERNAL_H
- /* Define DEBUG to get debugging output on stdout. */
+ /* Define PCRE_DEBUG to get debugging output on stdout. */
  #if 0
- #define DEBUG
+ #define PCRE_DEBUG
  #endif
- /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
+ /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
+ #ifndef COMPILE_PCRE16
+ #define COMPILE_PCRE8
+ #endif
+ /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
  script prevents both being selected, but not everybody uses "configure". */
- #if defined EBCDIC && defined SUPPORT_UTF8
+ #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
- #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
+ #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
  #endif
  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
  "configure" script ensures this, but not everybody uses "configure". */
  #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
-Line 74 
 It turns out that the Mac Debugging.h he
+Line 80 
 It turns out that the Mac Debugging.h he
  be absolutely sure we get our version. */
  #undef DPRINTF
- #ifdef DEBUG
+ #ifdef PCRE_DEBUG
  #define DPRINTF(p) printf p
  #else
  #define DPRINTF(p) /* Nothing */
-Line 86 
 setjmp and stdarg are used is when NO_RE
+Line 92 
 setjmp and stdarg are used is when NO_RE
  #include <ctype.h>
  #include <limits.h>
- #include <setjmp.h>
- #include <stdarg.h>
  #include <stddef.h>
  #include <stdio.h>
  #include <stdlib.h>
-Line 160 
 set, we ensure here that it has no effec
+Line 164 
 set, we ensure here that it has no effec
  #define PCRE_CALL_CONVENTION
  #endif
- /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
+ /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
  cannot determine these outside the compilation (e.g. by running a program as
  part of "configure") because PCRE is often cross-compiled for use on other
  systems. Instead we make use of the maximum sizes that are available at
  preprocessor time in standard C environments. */
+ typedef unsigned char pcre_uint8;
  #if USHRT_MAX == 65535
    typedef unsigned short pcre_uint16;
    typedef short pcre_int16;
-Line 186 
 preprocessor time in standard C environm
+Line 192 
 preprocessor time in standard C environm
    #error Cannot determine a type for 32-bit unsigned integers
  #endif
+ /* When checking for integer overflow in pcre_compile(), we need to handle
+ large integers. If a 64-bit integer type is available, we can use that.
+ Otherwise we have to cast to double, which of course requires floating point
+ arithmetic. Handle this by defining a macro for the appropriate type. If
+ stdint.h is available, include it; it may define INT64_MAX. Systems that do not
+ have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
+ by "configure". */
+ #if HAVE_STDINT_H
+ #include <stdint.h>
+ #elif HAVE_INTTYPES_H
+ #include <inttypes.h>
+ #endif
+ #if defined INT64_MAX || defined int64_t
+ #define INT64_OR_DOUBLE int64_t
+ #else
+ #define INT64_OR_DOUBLE double
+ #endif
  /* All character handling must be done as unsigned characters. Otherwise there
  are problems with top-bit-set characters and functions such as isspace().
- 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 *,
- 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
- 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
- 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
+ it is defined in sys/types, so use "uschar" instead. */
+ #ifdef COMPILE_PCRE8
+ typedef unsigned char pcre_uchar;
+ #define IN_UCHARS(x) (x)
+ #else
+ #ifdef COMPILE_PCRE16
+ #if USHRT_MAX != 65535
+ /* This is a warning message. Change PCRE_SCHAR16 to a 16 bit data type in
+ pcre.h(.in) and disable (comment out) this message. */
+ #error Warning: PCRE_SCHAR16 is not a 16 bit data type.
+ #endif
+ typedef pcre_uint16 pcre_uchar;
+ #define IN_UCHARS(x) ((x) << 1)
+ #else
+ #error Unsupported compiling mode
+ #endif /* COMPILE_PCRE16 */
- typedef unsigned char uschar;
+ #endif /* COMPILE_PCRE8 */
  /* This is an unsigned int value that no character can ever have. UTF-8
  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
-Line 216 
 start/end of string field names are. */
+Line 263 
 start/end of string field names are. */
  #define IS_NEWLINE(p) \
    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
      ((p) < NLBLOCK->PSEND && \
-      _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
+      PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
         utf8)) \
      : \
      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
-Line 230 
 start/end of string field names are. */
+Line 277 
 start/end of string field names are. */
  #define WAS_NEWLINE(p) \
    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
      ((p) > NLBLOCK->PSSTART && \
-      _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
+      PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
         &(NLBLOCK->nllen), utf8)) \
      : \
      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
-Line 249 
 used for the external interface and appe
+Line 296 
 used for the external interface and appe
  must begin with PCRE_. */
  #ifdef CUSTOM_SUBJECT_PTR
- #define PCRE_SPTR CUSTOM_SUBJECT_PTR
+ #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
- #define USPTR CUSTOM_SUBJECT_PTR
  #else
- #define PCRE_SPTR const char *
+ #define PCRE_PUCHAR const pcre_uchar *
- #define USPTR const unsigned char *
  #endif
  /* Include the public PCRE header and the definitions of UCP character property
  values. */
-Line 325 
 The macros are controlled by the value o
+Line 368 
 The macros are controlled by the value o
  the config.h file, but can be overridden by using -D on the command line. This
  is automated on Unix systems via the "configure" command. */
+ #ifdef COMPILE_PCRE8
  #if LINK_SIZE == 2
  #define PUT(a,n,d)   \
-Line 361 
 is automated on Unix systems via the "co
+Line 406 
 is automated on Unix systems via the "co
  #define GET(a,n) \
    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
- #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */
+ /* Keep it positive */
+ #define MAX_PATTERN_SIZE (1 << 30)
+ #else
+ #error LINK_SIZE must be either 2, 3, or 4
+ #endif
+ #else /* COMPILE_PCRE8 */
+ #ifdef COMPILE_PCRE16
+ #if LINK_SIZE == 2
+ #undef LINK_SIZE
+ #define LINK_SIZE 1
+ #define PUT(a,n,d)   \
+   (a[n] = (d))
+ #define GET(a,n) \
+   (a[n])
+ #define MAX_PATTERN_SIZE (1 << 16)
+ #elif LINK_SIZE == 3 || LINK_SIZE == 4
+ #undef LINK_SIZE
+ #define LINK_SIZE 2
+ #define PUT(a,n,d)   \
+   (a[n] = (d) >> 16), \
+   (a[(n)+1] = (d) & 65536)
+ #define GET(a,n) \
+   (((a)[n] << 16) | (a)[(n)+1])
+ /* Keep it positive */
+ #define MAX_PATTERN_SIZE (1 << 30)
  #else
  #error LINK_SIZE must be either 2, 3, or 4
  #endif
+ #else
+ #error Unsupported compiling mode
+ #endif /* COMPILE_PCRE16 */
+ #endif /* COMPILE_PCRE8 */
  /* Convenience macro defined in terms of the others */
-Line 378 
 is automated on Unix systems via the "co
+Line 464 
 is automated on Unix systems via the "co
  offsets changes. There are used for repeat counts and for other things such as
  capturing parenthesis numbers in back references. */
+ #ifdef COMPILE_PCRE8
+ #define IMM2_SIZE 2
  #define PUT2(a,n,d)   \
    a[n] = (d) >> 8; \
    a[(n)+1] = (d) & 255
-Line 385 
 capturing parenthesis numbers in back re
+Line 475 
 capturing parenthesis numbers in back re
  #define GET2(a,n) \
    (((a)[n] << 8) | (a)[(n)+1])
- #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2
+ #else /* COMPILE_PCRE8 */
+ #ifdef COMPILE_PCRE16
+ #define IMM2_SIZE 1
+ #define PUT2(a,n,d)   \
+    a[n] = d
+ #define GET2(a,n) \
+    a[n]
+ #else
+ #error Unsupported compiling mode
+ #endif /* COMPILE_PCRE16 */
+ #endif /* COMPILE_PCRE8 */
+ #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
  /* When UTF-8 encoding is being used, a character is no longer just a single
  byte. The macros for character handling generate simple sequences when used in
- byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
+ byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
- 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
- support is omitted, we don't even define it. */
+ never be called in byte mode. To make sure they can never even appear when
+ UTF-8 support is omitted, we don't even define them. */
  #ifndef SUPPORT_UTF8
  #define GETCHAR(c, eptr) c = *eptr;
-Line 400 
 support is omitted, we don't even define
+Line 508 
 support is omitted, we don't even define
  #define GETCHARINC(c, eptr) c = *eptr++;
  #define GETCHARINCTEST(c, eptr) c = *eptr++;
  #define GETCHARLEN(c, eptr, len) c = *eptr;
+ /* #define GETCHARLENTEST(c, eptr, len) */
  /* #define BACKCHAR(eptr) */
  #else   /* SUPPORT_UTF8 */
+ /* These macros were originally written in the form of loops that used data
+ from the tables whose names start with PRIV(utf8_table). They were rewritten by
+ a user so as not to use loops, because in some environments this gives a
+ significant performance advantage, and it seems never to do any harm. */
+ /* Base macro to pick up the remaining bytes of a UTF-8 character, not
+ advancing the pointer. */
+ #define GETUTF8(c, eptr) \
+     { \
+     if ((c & 0x20) == 0) \
+       c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
+     else if ((c & 0x10) == 0) \
+       c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
+     else if ((c & 0x08) == 0) \
+       c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
+       ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
+     else if ((c & 0x04) == 0) \
+       c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
+           ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
+           (eptr[4] & 0x3f); \
+     else \
+       c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
+           ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
+           ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
+     }
  /* Get the next UTF-8 character, not advancing the pointer. This is called when
  we know we are in UTF-8 mode. */
  #define GETCHAR(c, eptr) \
    c = *eptr; \
-   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; \
-       } \
-     }
  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
  pointer. */
  #define GETCHARTEST(c, eptr) \
    c = *eptr; \
-   if (utf8 && c >= 0xc0) \
+   if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
+ /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
+ the pointer. */
+ #define GETUTF8INC(c, eptr) \
      { \
-     int gcii; \
+     if ((c & 0x20) == 0) \
-     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
+       c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
-     int gcss = 6*gcaa; \
+     else if ((c & 0x10) == 0) \
-     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
-     for (gcii = 1; gcii <= gcaa; gcii++) \
        { \
-       gcss -= 6; \
+       c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
-       c |= (eptr[gcii] & 0x3f) << gcss; \
+       eptr += 2; \
+       } \
+     else if ((c & 0x08) == 0) \
+       { \
+       c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
+           ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
+       eptr += 3; \
+       } \
+     else if ((c & 0x04) == 0) \
+       { \
+       c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
+           ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
+           (eptr[3] & 0x3f); \
+       eptr += 4; \
+       } \
+     else \
+       { \
+       c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
+           ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
+           ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
+       eptr += 5; \
        } \
      }
-Line 445 
 know we are in UTF-8 mode. */
+Line 593 
 know we are in UTF-8 mode. */
  #define GETCHARINC(c, eptr) \
    c = *eptr++; \
-   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; \
-       } \
-     }
- /* 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.
+ This is called when we don't know if we are in UTF-8 mode. */
  #define GETCHARINCTEST(c, eptr) \
    c = *eptr++; \
-   if (utf8 && c >= 0xc0) \
+   if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
+ /* Base macro to pick up the remaining bytes of a UTF-8 character, not
+ advancing the pointer, incrementing the length. */
+ #define GETUTF8LEN(c, eptr, len) \
      { \
-     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
+     if ((c & 0x20) == 0) \
-     int gcss = 6*gcaa; \
+       { \
-     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
+       c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
-     while (gcaa-- > 0) \
+       len++; \
+       } \
+     else if ((c & 0x10)  == 0) \
+       { \
+       c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
+       len += 2; \
+       } \
+     else if ((c & 0x08)  == 0) \
+       {\
+       c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
+           ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
+       len += 3; \
+       } \
+     else if ((c & 0x04)  == 0) \
        { \
-       gcss -= 6; \
+       c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
-       c |= (*eptr++ & 0x3f) << gcss; \
+           ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
+           (eptr[4] & 0x3f); \
+       len += 4; \
+       } \
+     else \
+       {\
+       c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
+           ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
+           ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
+       len += 5; \
        } \
      }
-Line 478 
 if there are extra bytes. This is called
+Line 644 
 if there are extra bytes. This is called
  #define GETCHARLEN(c, eptr, len) \
    c = *eptr; \
-   if (c >= 0xc0) \
+   if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
-     { \
-     int gcii; \
-     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
-     int gcss = 6*gcaa; \
-     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
-     for (gcii = 1; gcii <= gcaa; gcii++) \
-       { \
-       gcss -= 6; \
-       c |= (eptr[gcii] & 0x3f) << gcss; \
-       } \
-     len += gcaa; \
-     }
  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
  pointer, incrementing length if there are extra bytes. This is called when we
- know we are in UTF-8 mode. */
+ do not know if we are in UTF-8 mode. */
  #define GETCHARLENTEST(c, eptr, len) \
    c = *eptr; \
-   if (utf8 && c >= 0xc0) \
+   if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
-     { \
-     int gcii; \
-     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
-     int gcss = 6*gcaa; \
-     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
-     for (gcii = 1; gcii <= gcaa; gcii++) \
-       { \
-       gcss -= 6; \
-       c |= (eptr[gcii] & 0x3f) << gcss; \
-       } \
-     len += gcaa; \
-     }
  /* If the pointer is not at the start of a character, move it back until
  it is. This is called only in UTF-8 mode - we don't put a test within the macro
-Line 518 
 because almost all calls are already wit
+Line 660 
 because almost all calls are already wit
  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
- #endif
+ #endif  /* SUPPORT_UTF8 */
  /* In case there is no definition of offsetof() provided - though any proper
-Line 529 
 Standard C system should have one. */
+Line 671 
 Standard C system should have one. */
  #endif
- /* These are the public options that can change during matching. */
- #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
  /* Private flags containing information about the compiled regex. They used to
  live at the top end of the options word, but that got almost full, so now they
- are in a 16-bit flags word. */
+ are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
+ the restrictions on partial matching have been lifted. It remains for backwards
+ compatibility. */
  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
-Line 543 
 are in a 16-bit flags word. */
+Line 683 
 are in a 16-bit flags word. */
  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
+ #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
- /* Options for the "extra" block produced by pcre_study(). */
+ /* Flags for the "extra" block produced by pcre_study(). */
- #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
+ #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
+ #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
  /* Masks for identifying the public options that are permitted at compile
  time, run time, or study time, respectively. */
-Line 559 
 time, run time, or study time, respectiv
+Line 701 
 time, run time, or study time, respectiv
     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
-    PCRE_JAVASCRIPT_COMPAT)
+    PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
  #define PUBLIC_EXEC_OPTIONS \
-   (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
+   (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
-    PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
+    PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
-    PCRE_NO_START_OPTIMIZE)
+    PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
  #define PUBLIC_DFA_EXEC_OPTIONS \
-   (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
+   (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
-    PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
+    PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
-    PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
+    PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
+    PCRE_NO_START_OPTIMIZE)
- #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
+ #define PUBLIC_STUDY_OPTIONS \
+    PCRE_STUDY_JIT_COMPILE
  /* Magic number to provide a small check against being handed junk. Also used
  to detect whether a pattern was compiled on a host of different endianness. */
-Line 853 
 so that PCRE works on both ASCII and EBC
+Line 997 
 so that PCRE works on both ASCII and EBC
  #define STRING_COMMIT0              "COMMIT\0"
  #define STRING_F0                   "F\0"
  #define STRING_FAIL0                "FAIL\0"
+ #define STRING_MARK0                "MARK\0"
  #define STRING_PRUNE0               "PRUNE\0"
  #define STRING_SKIP0                "SKIP\0"
  #define STRING_THEN                 "THEN"
-Line 874 
 so that PCRE works on both ASCII and EBC
+Line 1019 
 so that PCRE works on both ASCII and EBC
  #define STRING_DEFINE               "DEFINE"
- #define STRING_CR_RIGHTPAR          "CR)"
+ #define STRING_CR_RIGHTPAR             "CR)"
- #define STRING_LF_RIGHTPAR          "LF)"
+ #define STRING_LF_RIGHTPAR             "LF)"
- #define STRING_CRLF_RIGHTPAR        "CRLF)"
+ #define STRING_CRLF_RIGHTPAR           "CRLF)"
- #define STRING_ANY_RIGHTPAR         "ANY)"
+ #define STRING_ANY_RIGHTPAR            "ANY)"
- #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
+ #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
- #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
+ #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
- #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
+ #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
+ #define STRING_UTF8_RIGHTPAR           "UTF8)"
+ #define STRING_UCP_RIGHTPAR            "UCP)"
+ #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
  #else  /* SUPPORT_UTF8 */
-Line 1104 
 only. */
+Line 1252 
 only. */
  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
  #define STRING_F0                   STR_F "\0"
  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
+ #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
  #define STRING_THEN                 STR_T STR_H STR_E STR_N
-Line 1125 
 only. */
+Line 1274 
 only. */
  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
- #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
+ #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
- #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
+ #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
- #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
- #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
- #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
- #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
- #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
+ #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
+ #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
  #endif  /* SUPPORT_UTF8 */
-Line 1164 
 only. */
+Line 1316 
 only. */
  #define PT_ANY        0    /* Any property - matches all chars */
  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
- #define PT_GC         2    /* General characteristic (e.g. L) */
+ #define PT_GC         2    /* Specified general characteristic (e.g. L) */
- #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
+ #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
  #define PT_SC         4    /* Script (e.g. Han) */
+ #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
+ #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
+ #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
+ #define PT_WORD       8    /* Word - L plus N plus underscore */
  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
  contain UTF-8 characters with values greater than 255. */
-Line 1183 
 contain UTF-8 characters with values gre
+Line 1339 
 contain UTF-8 characters with values gre
  /* These are escaped items that aren't just an encoding of a particular data
  value such as \n. They must have non-zero values, as check_escape() returns
  their negation. Also, they must appear in the same order as in the opcode
- 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
- corresponds to "." rather than an escape sequence, and another for OP_ALLANY
+ corresponds to "." in DOTALL mode rather than an escape sequence. It is also
- (which is used for [^] in JavaScript compatibility mode).
+ used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
+ non-DOTALL mode, "." behaves like \N.
+ The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
+ when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
+ They must be contiguous, and remain in order so that the replacements can be
+ looked up from a table.
  The final escape must be ESC_REF as subsequent values are used for
  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
-Line 1195 
 put in between that don't consume a char
+Line 1357 
 put in between that don't consume a char
  */
  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
-        ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
+        ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
-        ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
+        ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
+        ESC_E, ESC_Q, ESC_g, ESC_k,
+        ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
         ESC_REF };
  /* 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.
  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
- 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
- "coptable" in pcre_dfa_exec.c that must be updated. */
+ "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
  enum {
    OP_END,            /* 0 End of pattern */
-Line 1223 
 enum {
+Line 1386 
 enum {
    OP_WHITESPACE,         /*  9 \s */
    OP_NOT_WORDCHAR,       /* 10 \W */
    OP_WORDCHAR,           /* 11 \w */
-   OP_ANY,            /* 12 Match any character (subject to DOTALL) */
-   OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
+   OP_ANY,            /* 12 Match any character except newline */
+   OP_ALLANY,         /* 13 Match any character */
    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
    OP_NOTPROP,        /* 15 \P (not Unicode property) */
    OP_PROP,           /* 16 \p (Unicode property) */
-Line 1237 
 enum {
+Line 1401 
 enum {
    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
    OP_EOD,            /* 24 End of data: \z */
-   OP_OPT,            /* 25 Set runtime options */
+   OP_CIRC,           /* 25 Start of line - not multiline */
-   OP_CIRC,           /* 26 Start of line - varies with multiline switch */
+   OP_CIRCM,          /* 26 Start of line - multiline */
-   OP_DOLL,           /* 27 End of line - varies with multiline switch */
+   OP_DOLL,           /* 27 End of line - not multiline */
-   OP_CHAR,           /* 28 Match one character, casefully */
+   OP_DOLLM,          /* 28 End of line - multiline */
-   OP_CHARNC,         /* 29 Match one character, caselessly */
+   OP_CHAR,           /* 29 Match one character, casefully */
-   OP_NOT,            /* 30 Match one character, not the following one */
+   OP_CHARI,          /* 30 Match one character, caselessly */
+   OP_NOT,            /* 31 Match one character, not the given one, casefully */
-   OP_STAR,           /* 31 The maximizing and minimizing versions of */
+   OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
-   OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
-   OP_PLUS,           /* 33 the minimizing one second. */
+   /* The following sets of 13 opcodes must always be kept in step because
-   OP_MINPLUS,        /* 34 This first set applies to single characters.*/
+   the offset from the first one is used to generate the others. */
-   OP_QUERY,          /* 35 */
-   OP_MINQUERY,       /* 36 */
+   /**** Single characters, caseful, must precede the caseless ones ****/
-   OP_UPTO,           /* 37 From 0 to n matches */
+   OP_STAR,           /* 33 The maximizing and minimizing versions of */
-   OP_MINUPTO,        /* 38 */
+   OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
-   OP_EXACT,          /* 39 Exactly n matches */
+   OP_PLUS,           /* 35 the minimizing one second. */
+   OP_MINPLUS,        /* 36 */
-   OP_POSSTAR,        /* 40 Possessified star */
+   OP_QUERY,          /* 37 */
-   OP_POSPLUS,        /* 41 Possessified plus */
+   OP_MINQUERY,       /* 38 */
-   OP_POSQUERY,       /* 42 Posesssified query */
-   OP_POSUPTO,        /* 43 Possessified upto */
+   OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
+   OP_MINUPTO,        /* 40 */
-   OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
+   OP_EXACT,          /* 41 Exactly n matches */
-   OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
-   OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
+   OP_POSSTAR,        /* 42 Possessified star, caseful */
-   OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
+   OP_POSPLUS,        /* 43 Possessified plus, caseful */
-   OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
+   OP_POSQUERY,       /* 44 Posesssified query, caseful */
-   OP_NOTMINQUERY,    /* 49 */
+   OP_POSUPTO,        /* 45 Possessified upto, caseful */
-   OP_NOTUPTO,        /* 50 From 0 to n matches */
+   /**** Single characters, caseless, must follow the caseful ones */
-   OP_NOTMINUPTO,     /* 51 */
-   OP_NOTEXACT,       /* 52 Exactly n matches */
+   OP_STARI,          /* 46 */
+   OP_MINSTARI,       /* 47 */
-   OP_NOTPOSSTAR,     /* 53 Possessified versions */
+   OP_PLUSI,          /* 48 */
-   OP_NOTPOSPLUS,     /* 54 */
+   OP_MINPLUSI,       /* 49 */
-   OP_NOTPOSQUERY,    /* 55 */
+   OP_QUERYI,         /* 50 */
-   OP_NOTPOSUPTO,     /* 56 */
+   OP_MINQUERYI,      /* 51 */
-   OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
+   OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
-   OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
+   OP_MINUPTOI,       /* 53 */
-   OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
+   OP_EXACTI,         /* 54 */
-   OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
-   OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
+   OP_POSSTARI,       /* 55 Possessified star, caseless */
-   OP_TYPEMINQUERY,   /* 62 */
+   OP_POSPLUSI,       /* 56 Possessified plus, caseless */
+   OP_POSQUERYI,      /* 57 Posesssified query, caseless */
-   OP_TYPEUPTO,       /* 63 From 0 to n matches */
+   OP_POSUPTOI,       /* 58 Possessified upto, caseless */
-   OP_TYPEMINUPTO,    /* 64 */
-   OP_TYPEEXACT,      /* 65 Exactly n matches */
+   /**** The negated ones must follow the non-negated ones, and match them ****/
+   /**** Negated single character, caseful; must precede the caseless ones ****/
-   OP_TYPEPOSSTAR,    /* 66 Possessified versions */
-   OP_TYPEPOSPLUS,    /* 67 */
+   OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
-   OP_TYPEPOSQUERY,   /* 68 */
+   OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
-   OP_TYPEPOSUPTO,    /* 69 */
+   OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
+   OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
-   OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
+   OP_NOTQUERY,       /* 63 */
-   OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
+   OP_NOTMINQUERY,    /* 64 */
-   OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
-   OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
+   OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
-   OP_CRQUERY,        /* 74 These are for character classes and back refs */
+   OP_NOTMINUPTO,     /* 66 */
-   OP_CRMINQUERY,     /* 75 */
+   OP_NOTEXACT,       /* 67 Exactly n matches */
-   OP_CRRANGE,        /* 76 These are different to the three sets above. */
-   OP_CRMINRANGE,     /* 77 */
+   OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
+   OP_NOTPOSPLUS,     /* 69 */
-   OP_CLASS,          /* 78 Match a character class, chars < 256 only */
+   OP_NOTPOSQUERY,    /* 70 */
-   OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
+   OP_NOTPOSUPTO,     /* 71 */
-                            class - the difference is relevant only when a UTF-8
-                            character > 255 is encountered. */
+   /**** Negated single character, caseless; must follow the caseful ones ****/
-   OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
+   OP_NOTSTARI,       /* 72 */
-                            class. This does both positive and negative. */
+   OP_NOTMINSTARI,    /* 73 */
+   OP_NOTPLUSI,       /* 74 */
-   OP_REF,            /* 81 Match a back reference */
+   OP_NOTMINPLUSI,    /* 75 */
-   OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
+   OP_NOTQUERYI,      /* 76 */
-   OP_CALLOUT,        /* 83 Call out to external function if provided */
+   OP_NOTMINQUERYI,   /* 77 */
-   OP_ALT,            /* 84 Start of alternation */
+   OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
-   OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
+   OP_NOTMINUPTOI,    /* 79 */
-   OP_KETRMAX,        /* 86 These two must remain together and in this */
+   OP_NOTEXACTI,      /* 80 Exactly n matches */
-   OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
+   OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
-   /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
+   OP_NOTPOSPLUSI,    /* 82 */
+   OP_NOTPOSQUERYI,   /* 83 */
-   OP_ASSERT,         /* 88 Positive lookahead */
+   OP_NOTPOSUPTOI,    /* 84 */
-   OP_ASSERT_NOT,     /* 89 Negative lookahead */
-   OP_ASSERTBACK,     /* 90 Positive lookbehind */
+   /**** Character types ****/
-   OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
-   OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
+   OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
+   OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
-   /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
+   OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
-   as there's a test for >= ONCE for a subpattern that isn't an assertion. */
+   OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
+   OP_TYPEQUERY,      /* 89 */
-   OP_ONCE,           /* 93 Atomic group */
+   OP_TYPEMINQUERY,   /* 90 */
-   OP_BRA,            /* 94 Start of non-capturing bracket */
-   OP_CBRA,           /* 95 Start of capturing bracket */
+   OP_TYPEUPTO,       /* 91 From 0 to n matches */
-   OP_COND,           /* 96 Conditional group */
+   OP_TYPEMINUPTO,    /* 92 */
+   OP_TYPEEXACT,      /* 93 Exactly n matches */
+   OP_TYPEPOSSTAR,    /* 94 Possessified versions */
+   OP_TYPEPOSPLUS,    /* 95 */
+   OP_TYPEPOSQUERY,   /* 96 */
+   OP_TYPEPOSUPTO,    /* 97 */
+   /* These are used for character classes and back references; only the
+   first six are the same as the sets above. */
+   OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
+   OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
+   OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
+   OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
+   OP_CRQUERY,        /* 102 */
+   OP_CRMINQUERY,     /* 103 */
+   OP_CRRANGE,        /* 104 These are different to the three sets above. */
+   OP_CRMINRANGE,     /* 105 */
+   /* End of quantifier opcodes */
+   OP_CLASS,          /* 106 Match a character class, chars < 256 only */
+   OP_NCLASS,         /* 107 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,         /* 108 Extended class for handling UTF-8 chars within the
+                               class. This does both positive and negative. */
+   OP_REF,            /* 109 Match a back reference, casefully */
+   OP_REFI,           /* 110 Match a back reference, caselessly */
+   OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
+   OP_CALLOUT,        /* 112 Call out to external function if provided */
+   OP_ALT,            /* 113 Start of alternation */
+   OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
+   OP_KETRMAX,        /* 115 These two must remain together and in this */
+   OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
+   OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
+   /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
+   asserts must remain in order. */
+   OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
+   OP_ASSERT,         /* 119 Positive lookahead */
+   OP_ASSERT_NOT,     /* 120 Negative lookahead */
+   OP_ASSERTBACK,     /* 121 Positive lookbehind */
+   OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
+   /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
+   after the assertions, with ONCE first, as there's a test for >= ONCE for a
+   subpattern that isn't an assertion. The POS versions must immediately follow
+   the non-POS versions in each case. */
+   OP_ONCE,           /* 123 Atomic group, contains captures */
+   OP_ONCE_NC,        /* 124 Atomic group containing no captures */
+   OP_BRA,            /* 125 Start of non-capturing bracket */
+   OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
+   OP_CBRA,           /* 127 Start of capturing bracket */
+   OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
+   OP_COND,           /* 129 Conditional group */
-   /* These three must follow the previous three, in the same order. There's a
+   /* These five must follow the previous five, in the same order. There's a
    check for >= SBRA to distinguish the two sets. */
-   OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
+   OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
-   OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
+   OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
-   OP_SCOND,          /* 99 Conditional group, check empty */
+   OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
+   OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
-   OP_CREF,           /* 100 Used to hold a capture number as condition */
+   OP_SCOND,          /* 134 Conditional group, check empty */
-   OP_RREF,           /* 101 Used to hold a recursion number as condition */
-   OP_DEF,            /* 102 The DEFINE condition */
+   /* The next two pairs must (respectively) be kept together. */
-   OP_BRAZERO,        /* 103 These two must remain together and in this */
+   OP_CREF,           /* 135 Used to hold a capture number as condition */
-   OP_BRAMINZERO,     /* 104 order. */
+   OP_NCREF,          /* 136 Same, but generated by a name reference*/
+   OP_RREF,           /* 137 Used to hold a recursion number as condition */
+   OP_NRREF,          /* 138 Same, but generated by a name reference*/
+   OP_DEF,            /* 139 The DEFINE condition */
+   OP_BRAZERO,        /* 140 These two must remain together and in this */
+   OP_BRAMINZERO,     /* 141 order. */
+   OP_BRAPOSZERO,     /* 142 */
    /* These are backtracking control verbs */
-   OP_PRUNE,          /* 105 */
+   OP_MARK,           /* 143 always has an argument */
-   OP_SKIP,           /* 106 */
+   OP_PRUNE,          /* 144 */
-   OP_THEN,           /* 107 */
+   OP_PRUNE_ARG,      /* 145 same, but with argument */
-   OP_COMMIT,         /* 108 */
+   OP_SKIP,           /* 146 */
+   OP_SKIP_ARG,       /* 147 same, but with argument */
+   OP_THEN,           /* 148 */
+   OP_THEN_ARG,       /* 149 same, but with argument */
+   OP_COMMIT,         /* 150 */
    /* These are forced failure and success verbs */
-   OP_FAIL,           /* 109 */
+   OP_FAIL,           /* 151 */
-   OP_ACCEPT,         /* 110 */
+   OP_ACCEPT,         /* 152 */
+   OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
+   OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
    /* This is used to skip a subpattern with a {0} quantifier */
-   OP_SKIPZERO        /* 111 */
+   OP_SKIPZERO,       /* 155 */
+   /* This is not an opcode, but is used to check that tables indexed by opcode
+   are the correct length, in order to catch updating errors - there have been
+   some in the past. */
+   OP_TABLE_LENGTH
  };
+ /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
+ definitions that follow must also be updated to match. There are also tables
+ called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
  /* This macro defines textual names for all the opcodes. These are used only
- 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
+ only in pcre_printint.c, which fills out the full names in many cases (and in
+ some cases doesn't actually use these names at all). */
  #define OP_NAME_LIST \
    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
    "extuni",  "\\Z", "\\z",                                        \
-   "Opt", "^", "$", "char", "charnc", "not",                       \
+   "^", "^", "$", "$", "char", "chari", "not", "noti",             \
-   "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
+   "*", "*?", "+", "+?", "?", "??",                                \
+   "{", "{", "{",                                                  \
    "*+","++", "?+", "{",                                           \
-   "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
+   "*", "*?", "+", "+?", "?", "??",                                \
+   "{", "{", "{",                                                  \
+   "*+","++", "?+", "{",                                           \
+   "*", "*?", "+", "+?", "?", "??",                                \
+   "{", "{", "{",                                                  \
+   "*+","++", "?+", "{",                                           \
+   "*", "*?", "+", "+?", "?", "??",                                \
+   "{", "{", "{",                                                  \
    "*+","++", "?+", "{",                                           \
    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
    "*+","++", "?+", "{",                                           \
    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
-   "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
+   "class", "nclass", "xclass", "Ref", "Refi",                     \
-   "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
+   "Recurse", "Callout",                                           \
-   "AssertB", "AssertB not", "Reverse",                            \
+   "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
-   "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
+   "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
-   "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
+   "Once", "Once_NC",                                              \
-   "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
+   "Bra", "BraPos", "CBra", "CBraPos",                             \
-   "Skip zero"
+   "Cond",                                                         \
+   "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
+   "SCond",                                                        \
+   "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
+   "Brazero", "Braminzero", "Braposzero",                          \
+   "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
+   "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
+   "*ACCEPT", "*ASSERT_ACCEPT",                                    \
+   "Close", "Skip zero"
  /* This macro defines the length of fixed length operations in the compiled
-Line 1405 
 in UTF-8 mode. The code that uses this t
+Line 1670 
 in UTF-8 mode. The code that uses this t
 , 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
 , 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
 , 1, 1,                       /* Any, AllAny, Anybyte                   */ \
-, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
+, 3,                          /* \P, \p                                 */ \
 , 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
-, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
+,                             /* \X                                     */ \
+, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
 ,                             /* Char  - the minimum length             */ \
-,                             /* Charnc  - the minimum length           */ \
+,                             /* Chari  - the minimum length            */ \
 ,                             /* not                                    */ \
-   /* Positive single-char repeats                            ** These are  */ \
+,                             /* noti                                   */ \
-, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
+   /* Positive single-char repeats                             ** These are */ \
-, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
+, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
-, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
++IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
++IMM2_SIZE,                   /* exact                                  */ \
+, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
+, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
++IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
++IMM2_SIZE,                   /* exact I                                */ \
+, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
    /* Negative single-char repeats - only for chars < 256                   */ \
 , 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
-, 4, 4,                       /* NOT upto, minupto, exact               */ \
++IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
-, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
++IMM2_SIZE,                   /* NOT exact                              */ \
+, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
+, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
++IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
++IMM2_SIZE,                   /* NOT exact I                            */ \
+, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
    /* Positive type repeats                                                 */ \
 , 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
-, 4, 4,                       /* Type upto, minupto, exact              */ \
++IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
-, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
++IMM2_SIZE,                   /* Type exact                             */ \
+, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
    /* Character class & ref repeats                                         */ \
 , 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
-, 5,                          /* CRRANGE, CRMINRANGE                    */ \
++2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
 ,                             /* CLASS                                  */ \
 ,                             /* NCLASS                                 */ \
 ,                             /* XCLASS - variable length               */ \
-,                             /* REF                                    */ \
++IMM2_SIZE,                   /* REF                                    */ \
++IMM2_SIZE,                   /* REFI                                   */ \
 +LINK_SIZE,                   /* RECURSE                                */ \
 +2*LINK_SIZE,                 /* CALLOUT                                */ \
 +LINK_SIZE,                   /* Alt                                    */ \
 +LINK_SIZE,                   /* Ket                                    */ \
 +LINK_SIZE,                   /* KetRmax                                */ \
 +LINK_SIZE,                   /* KetRmin                                */ \
++LINK_SIZE,                   /* KetRpos                                */ \
++LINK_SIZE,                   /* Reverse                                */ \
 +LINK_SIZE,                   /* Assert                                 */ \
 +LINK_SIZE,                   /* Assert not                             */ \
 +LINK_SIZE,                   /* Assert behind                          */ \
 +LINK_SIZE,                   /* Assert behind not                      */ \
-+LINK_SIZE,                   /* Reverse                                */ \
 +LINK_SIZE,                   /* ONCE                                   */ \
++LINK_SIZE,                   /* ONCE_NC                                */ \
 +LINK_SIZE,                   /* BRA                                    */ \
-+LINK_SIZE,                   /* CBRA                                   */ \
++LINK_SIZE,                   /* BRAPOS                                 */ \
++LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
++LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
 +LINK_SIZE,                   /* COND                                   */ \
 +LINK_SIZE,                   /* SBRA                                   */ \
-+LINK_SIZE,                   /* SCBRA                                  */ \
++LINK_SIZE,                   /* SBRAPOS                                */ \
++LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
++LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
 +LINK_SIZE,                   /* SCOND                                  */ \
-,                             /* CREF                                   */ \
++IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
-,                             /* RREF                                   */ \
++IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
 ,                             /* DEF                                    */ \
-, 1,                          /* BRAZERO, BRAMINZERO                    */ \
+, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
-, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
+, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
-, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
+, 3,                          /* SKIP, SKIP_ARG                         */ \
+, 3,                          /* THEN, THEN_ARG                         */ \
+, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
++IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
- /* 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"
+ condition. */
  #define RREF_ANY  0xffff
- /* 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
- tracked. */
+ easily be tracked. When a new number is added, the table called eint in
+ pcreposix.c must be updated. */
  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
-Line 1469 
 enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
+Line 1758 
 enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
-        ERR60, ERR61, ERR62, ERR63, ERR64 };
+        ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
+        ERR70, ERRCOUNT };
  /* The real format of the start of the pcre block; the index of names and the
  code vector run on as long as necessary after the end. We store an explicit
-Line 1485 
 Because people can now save and re-use c
+Line 1775 
 Because people can now save and re-use c
  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:
+ NOTE NOTE NOTE
  */
  typedef struct real_pcre {
-Line 1503 
 typedef struct real_pcre {
+Line 1793 
 typedef struct real_pcre {
    pcre_uint16 name_count;         /* Number of name items */
    pcre_uint16 ref_count;          /* Reference count */
-   const unsigned char *tables;    /* Pointer to tables or NULL for std */
+   const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
-   const unsigned char *nullpad;   /* NULL padding */
+   const pcre_uint8 *nullpad;      /* NULL padding */
  } real_pcre;
  /* The format of the block used to store data from pcre_study(). The same
-Line 1512 
 remark (see NOTE above) about extending
+Line 1802 
 remark (see NOTE above) about extending
  typedef struct pcre_study_data {
    pcre_uint32 size;               /* Total that was malloced */
-   pcre_uint32 options;
+   pcre_uint32 flags;              /* Private flags */
-   uschar start_bits[32];
+   pcre_uint8 start_bits[32];      /* Starting char bits */
+   pcre_uint32 minlength;          /* Minimum subject length */
  } pcre_study_data;
+ /* Structure for building a chain of open capturing subpatterns during
+ compiling, so that instructions to close them can be compiled when (*ACCEPT) is
+ encountered. This is also used to identify subpatterns that contain recursive
+ back references to themselves, so that they can be made atomic. */
+ typedef struct open_capitem {
+   struct open_capitem *next;    /* Chain link */
+   pcre_uint16 number;           /* Capture number */
+   pcre_uint16 flag;             /* Set TRUE if recursive back ref */
+ } open_capitem;
  /* Structure for passing "static" information around between the functions
  doing the compiling, so that they are thread-safe. */
  typedef struct compile_data {
-   const uschar *lcc;            /* Points to lower casing table */
+   const pcre_uint8 *lcc;            /* Points to lower casing table */
-   const uschar *fcc;            /* Points to case-flipping table */
+   const pcre_uint8 *fcc;            /* Points to case-flipping table */
-   const uschar *cbits;          /* Points to character type table */
+   const pcre_uint8 *cbits;          /* Points to character type table */
-   const uschar *ctypes;         /* Points to table of type maps */
+   const pcre_uint8 *ctypes;         /* Points to table of type maps */
-   const uschar *start_workspace;/* The start of working space */
+   const pcre_uchar *start_workspace;/* The start of working space */
-   const uschar *start_code;     /* The start of the compiled code */
+   const pcre_uchar *start_code;     /* The start of the compiled code */
-   const uschar *start_pattern;  /* The start of the pattern */
+   const pcre_uchar *start_pattern;  /* The start of the pattern */
-   const uschar *end_pattern;    /* The end of the pattern */
+   const pcre_uchar *end_pattern;    /* The end of the pattern */
-   uschar *hwm;                  /* High watermark of workspace */
+   open_capitem *open_caps;          /* Chain of open capture items */
-   uschar *name_table;           /* The name/number table */
+   pcre_uchar *hwm;                  /* High watermark of workspace */
-   int  names_found;             /* Number of entries so far */
+   pcre_uchar *name_table;           /* The name/number table */
-   int  name_entry_size;         /* Size of each entry */
+   int  names_found;                 /* Number of entries so far */
-   int  bracount;                /* Count of capturing parens as we compile */
+   int  name_entry_size;             /* Size of each entry */
-   int  final_bracount;          /* Saved value after first pass */
+   int  bracount;                    /* Count of capturing parens as we compile */
-   int  top_backref;             /* Maximum back reference */
+   int  final_bracount;              /* Saved value after first pass */
-   unsigned int backref_map;     /* Bitmap of low back refs */
+   int  top_backref;                 /* Maximum back reference */
-   int  external_options;        /* External (initial) options */
+   unsigned int backref_map;         /* Bitmap of low back refs */
-   int  external_flags;          /* External flag bits to be set */
+   int  assert_depth;                /* Depth of nested assertions */
-   int  req_varyopt;             /* "After variable item" flag for reqbyte */
+   int  external_options;            /* External (initial) options */
-   BOOL had_accept;              /* (*ACCEPT) encountered */
+   int  external_flags;              /* External flag bits to be set */
-   int  nltype;                  /* Newline type */
+   int  req_varyopt;                 /* "After variable item" flag for reqbyte */
-   int  nllen;                   /* Newline string length */
+   BOOL had_accept;                  /* (*ACCEPT) encountered */
-   uschar nl[4];                 /* Newline string when fixed length */
+   BOOL check_lookbehind;            /* Lookbehinds need later checking */
+   int  nltype;                      /* Newline type */
+   int  nllen;                       /* Newline string length */
+   pcre_uchar nl[4];                 /* Newline string when fixed length */
  } compile_data;
  /* Structure for maintaining a chain of pointers to the currently incomplete
- branches, for testing for left recursion. */
+ branches, for testing for left recursion while compiling. */
  typedef struct branch_chain {
    struct branch_chain *outer;
-   uschar *current;
+   pcre_uchar *current_branch;
  } branch_chain;
  /* Structure for items in a linked list that represents an explicit recursive
- call within the pattern. */
+ call within the pattern; used by pcre_exec(). */
  typedef struct recursion_info {
    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
-   int group_num;                /* Number of group that was called */
+   int group_num;                  /* Number of group that was called */
-   const uschar *after_call;     /* "Return value": points after the call in the expr */
+   int *offset_save;               /* Pointer to start of saved offsets */
-   USPTR save_start;             /* Old value of mstart */
+   int saved_max;                  /* Number of saved offsets */
-   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 */
  } recursion_info;
+ /* A similar structure for pcre_dfa_exec(). */
+ typedef struct dfa_recursion_info {
+   struct dfa_recursion_info *prevrec;
+   int group_num;
+   PCRE_PUCHAR subject_position;
+ } dfa_recursion_info;
  /* Structure for building a chain of data for holding the values of the subject
  pointer at the start of each subpattern, so as to detect when an empty string
- has been matched by a subpattern - to break infinite loops. */
+ has been matched by a subpattern - to break infinite loops; used by
+ pcre_exec(). */
  typedef struct eptrblock {
    struct eptrblock *epb_prev;
-   USPTR epb_saved_eptr;
+   PCRE_PUCHAR epb_saved_eptr;
  } eptrblock;
-Line 1582 
 typedef struct match_data {
+Line 1895 
 typedef struct match_data {
    unsigned long int match_call_count;      /* As it says */
    unsigned long int match_limit;           /* As it says */
    unsigned long int match_limit_recursion; /* As it says */
-   int   *offset_vector;         /* Offset vector */
+   int   *offset_vector;           /* Offset vector */
-   int    offset_end;            /* One past the end */
+   int    offset_end;              /* One past the end */
-   int    offset_max;            /* The maximum usable for return data */
+   int    offset_max;              /* The maximum usable for return data */
-   int    nltype;                /* Newline type */
+   int    nltype;                  /* Newline type */
-   int    nllen;                 /* Newline string length */
+   int    nllen;                   /* Newline string length */
-   uschar nl[4];                 /* Newline string when fixed */
+   int    name_count;              /* Number of names in name table */
-   const uschar *lcc;            /* Points to lower casing table */
+   int    name_entry_size;         /* Size of entry in names table */
-   const uschar *ctypes;         /* Points to table of type maps */
+   pcre_uchar *name_table;         /* Table of names */
-   BOOL   offset_overflow;       /* Set if too many extractions */
+   pcre_uchar nl[4];               /* Newline string when fixed */
-   BOOL   notbol;                /* NOTBOL flag */
+   const  pcre_uint8 *lcc;         /* Points to lower casing table */
-   BOOL   noteol;                /* NOTEOL flag */
+   const  pcre_uint8 *ctypes;      /* Points to table of type maps */
-   BOOL   utf8;                  /* UTF8 flag */
+   BOOL   offset_overflow;         /* Set if too many extractions */
-   BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
+   BOOL   notbol;                  /* NOTBOL flag */
-   BOOL   endonly;               /* Dollar not before final \n */
+   BOOL   noteol;                  /* NOTEOL flag */
-   BOOL   notempty;              /* Empty string match not wanted */
+   BOOL   utf8;                    /* UTF8 flag */
-   BOOL   partial;               /* PARTIAL flag */
+   BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
-   BOOL   hitend;                /* Hit the end of the subject at some point */
+   BOOL   use_ucp;                 /* PCRE_UCP flag */
-   BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
+   BOOL   endonly;                 /* Dollar not before final \n */
-   const uschar *start_code;     /* For use when recursing */
+   BOOL   notempty;                /* Empty string match not wanted */
-   USPTR  start_subject;         /* Start of the subject string */
+   BOOL   notempty_atstart;        /* Empty string match at start not wanted */
-   USPTR  end_subject;           /* End of the subject string */
+   BOOL   hitend;                  /* Hit the end of the subject at some point */
-   USPTR  start_match_ptr;       /* Start of matched string */
+   BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
-   USPTR  end_match_ptr;         /* Subject position at end match */
+   BOOL   hasthen;                 /* Pattern contains (*THEN) */
-   int    end_offset_top;        /* Highwater mark at end of match */
+   const  pcre_uchar *start_code;  /* For use when recursing */
-   int    capture_last;          /* Most recent capture number */
+   PCRE_PUCHAR start_subject;      /* Start of the subject string */
-   int    start_offset;          /* The start offset value */
+   PCRE_PUCHAR end_subject;        /* End of the subject string */
-   eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
+   PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
-   int    eptrn;                 /* Next free eptrblock */
+   PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
-   recursion_info *recursive;    /* Linked list of recursion data */
+   PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
-   void  *callout_data;          /* To pass back to callouts */
+   int    partial;                 /* PARTIAL options */
+   int    end_offset_top;          /* Highwater mark at end of match */
+   int    capture_last;            /* Most recent capture number */
+   int    start_offset;            /* The start offset value */
+   int    match_function_type;     /* Set for certain special calls of MATCH() */
+   eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
+   int    eptrn;                   /* Next free eptrblock */
+   recursion_info *recursive;      /* Linked list of recursion data */
+   void  *callout_data;            /* To pass back to callouts */
+   const  pcre_uchar *mark;        /* Mark pointer to pass back */
+   const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
  } match_data;
  /* A similar structure is used for the same purpose by the DFA matching
  functions. */
  typedef struct dfa_match_data {
-   const uschar *start_code;     /* Start of the compiled pattern */
+   const pcre_uchar *start_code;     /* Start of the compiled pattern */
-   const uschar *start_subject;  /* Start of the subject string */
+   const pcre_uchar *start_subject ; /* Start of the subject string */
-   const uschar *end_subject;    /* End of subject string */
+   const pcre_uchar *end_subject;    /* End of subject string */
-   const uschar *tables;         /* Character tables */
+   const pcre_uchar *start_used_ptr; /* Earliest consulted character */
-   int   moptions;               /* Match options */
+   const pcre_uint8 *tables;         /* Character tables */
-   int   poptions;               /* Pattern options */
+   int   start_offset;               /* The start offset value */
-   int    nltype;                /* Newline type */
+   int   moptions;                   /* Match options */
-   int    nllen;                 /* Newline string length */
+   int   poptions;                   /* Pattern options */
-   uschar nl[4];                 /* Newline string when fixed */
+   int   nltype;                     /* Newline type */
-   void  *callout_data;          /* To pass back to callouts */
+   int   nllen;                      /* Newline string length */
+   pcre_uchar nl[4];                 /* Newline string when fixed */
+   void *callout_data;               /* To pass back to callouts */
+   dfa_recursion_info *recursive;    /* Linked list of recursion data */
  } dfa_match_data;
  /* Bit definitions for entries in the pcre_ctypes table. */
-Line 1663 
 total length. */
+Line 1989 
 total length. */
  #define ctypes_offset (cbits_offset + cbit_length)
  #define tables_length (ctypes_offset + 256)
+ /* Internal function prefix */
+ #ifdef COMPILE_PCRE8
+ #define PRIV(name) _pcre_##name
+ #else
+ #ifdef COMPILE_PCRE16
+ #define PRIV(name) _pcre16_##name
+ #else
+ #error Unsupported compiling mode
+ #endif /* COMPILE_PCRE16 */
+ #endif /* COMPILE_PCRE8 */
  /* Layout of the UCP type table that translates property names into types and
  codes. Each entry used to point directly to a name, but to reduce the number of
  relocations in shared libraries, it now has an offset into a single string
-Line 1680 
 of the exported public functions. They h
+Line 2018 
 of the exported public functions. They h
  but are not part of the PCRE public API. The data for these tables is in the
  pcre_tables.c module. */
- extern const int    _pcre_utf8_table1[];
+ extern const int            PRIV(utf8_table1)[];
- extern const int    _pcre_utf8_table2[];
+ extern const int            PRIV(utf8_table2)[];
- extern const int    _pcre_utf8_table3[];
+ extern const int            PRIV(utf8_table3)[];
- extern const uschar _pcre_utf8_table4[];
+ extern const pcre_uint8     PRIV(utf8_table4)[];
- extern const int    _pcre_utf8_table1_size;
+ extern const int            PRIV(utf8_table1_size);
- extern const char   _pcre_utt_names[];
+ extern const char           PRIV(utt_names)[];
- extern const ucp_type_table _pcre_utt[];
+ extern const ucp_type_table PRIV(utt)[];
- extern const int _pcre_utt_size;
+ extern const int            PRIV(utt_size);
- extern const uschar _pcre_default_tables[];
+ extern const pcre_uint8     PRIV(default_tables)[];
- extern const uschar _pcre_OP_lengths[];
+ extern const pcre_uint8     PRIV(OP_lengths)[];
  /* Internal shared functions. These are functions that are used by more than
  one of the exported public functions. They have to be "external" in the C
  sense, but are not part of the PCRE public API. */
- extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
+ /* String comparison functions. */
-                       int *, BOOL);
+ #ifdef COMPILE_PCRE8
- extern int          _pcre_ord2utf8(int, uschar *);
- extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
-                       const pcre_study_data *, pcre_study_data *);
- extern int          _pcre_valid_utf8(const uschar *, int);
- extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
-                       int *, BOOL);
- extern BOOL         _pcre_xclass(int, const uschar *);
+ #define STRCMP_UC_UC(str1, str2) \
+   strcmp((char *)(str1), (char *)(str2))
+ #define STRCMP_UC_C8(str1, str2) \
+   strcmp((char *)(str1), (str2))
+ #define STRNCMP_UC_UC(str1, str2, num) \
+   strncmp((char *)(str1), (char *)(str2), (num))
+ #define STRNCMP_UC_C8(str1, str2, num) \
+   strncmp((char *)(str1), (str2), (num))
+ #define STRLEN_UC(str) strlen((const char *)str)
+ #else
+ extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
+                            const pcre_uchar *);
+ extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
+                            const char *);
+ extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
+                            const pcre_uchar *, unsigned int num);
+ extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
+                            const char *, unsigned int num);
+ extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
+ #define STRCMP_UC_UC(str1, str2) \
+   PRIV(strcmp_uc_uc)((str1), (str2))
+ #define STRCMP_UC_C8(str1, str2) \
+   PRIV(strcmp_uc_c8)((str1), (str2))
+ #define STRNCMP_UC_UC(str1, str2, num) \
+   PRIV(strncmp_uc_uc)((str1), (str2), (num))
+ #define STRNCMP_UC_C8(str1, str2, num) \
+   PRIV(strncmp_uc_c8)((str1), (str2), (num))
+ #define STRLEN_UC(str) PRIV(strlen_uc)(str)
+ #endif /* COMPILE_PCRE8 */
+ extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
+ extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
+                            int *, BOOL);
+ extern int               PRIV(ord2utf8)(int, pcre_uint8 *);
+ extern real_pcre        *PRIV(try_flipped)(const real_pcre *, real_pcre *,
+                            const pcre_study_data *, pcre_study_data *);
+ #ifndef COMPILE_PCRE16
+ extern int               PRIV(valid_utf8)(PCRE_PUCHAR, int, int *);
+ #else
+ extern int               PRIV(valid_utf16)(PCRE_PUCHAR, int, int *);
+ #endif
+ extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
+                            int *, BOOL);
+ extern BOOL              PRIV(xclass)(int, const pcre_uchar *);
+ #ifdef SUPPORT_JIT
+ extern void              PRIV(jit_compile)(const real_pcre *, pcre_extra *);
+ extern int               PRIV(jit_exec)(const real_pcre *, void *,
+                            const pcre_uchar *, int, int, int, int, int *, int);
+ extern void              PRIV(jit_free)(void *);
+ #endif
  /* Unicode character database (UCD) */
  typedef struct {
-   uschar script;
+   pcre_uint8 script;
-   uschar chartype;
+   pcre_uint8 chartype;
    pcre_int32 other_case;
  } ucd_record;
- extern const ucd_record  _pcre_ucd_records[];
+ extern const ucd_record  PRIV(ucd_records)[];
- extern const uschar      _pcre_ucd_stage1[];
+ extern const pcre_uint8  PRIV(ucd_stage1)[];
- extern const pcre_uint16 _pcre_ucd_stage2[];
+ extern const pcre_uint16 PRIV(ucd_stage2)[];
- extern const int         _pcre_ucp_gentype[];
+ extern const int         PRIV(ucp_gentype)[];
+ #ifdef SUPPORT_JIT
+ extern const int         PRIV(ucp_typerange)[];
+ #endif
  /* UCD access macros */
  #define UCD_BLOCK_SIZE 128
- #define GET_UCD(ch) (_pcre_ucd_records + \
+ #define GET_UCD(ch) (PRIV(ucd_records) + \
-         _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
+         PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
-         UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
+         UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
- #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
+ #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
  #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
  #endif

 Legend:



Removed from v.393
 


changed lines


 
Added in v.769
 Legend:



Removed from v.393
 


changed lines


 
Added in v.769
-Removed from v.393
+Added in v.769

webmaster@exim.org	ViewVC Help
Powered by ViewVC 1.1.12