Added Unicode ctype support.

configure.in

@@ -761,8 +761,7 @@ AC_CHECK_HEADERS(\
 	sys/vfs.h \
 	sys/vm86.h \
 	syscall.h \
-	ucontext.h \
-	wctype.h \
+	ucontext.h
 )
 AC_HEADER_STAT()
 

dlls/crtdll/mbstring.c

@@ -4,17 +4,8 @@
  * Copyright 1999 Alexandre Julliard
  */
 
-#include "config.h"
-
-#include <ctype.h>
-#include <limits.h>
-#include <stdlib.h>
-#include <string.h>
-#ifdef HAVE_WCTYPE_H
-#include <wctype.h>
-#endif
-
 #include "windef.h"
+#include "winbase.h"
 #include "crtdll.h"
 
 
@@ -23,9 +14,8 @@
  */
 LPSTR __cdecl CRTDLL__mbsinc( LPCSTR str )
 {
-    int len = mblen( str, MB_LEN_MAX );
-    if (len < 1) len = 1;
-    return (LPSTR)(str + len);
+    if (IsDBCSLeadByte( *str )) str++;
+    return (LPSTR)(str + 1);
 }
 
 
@@ -34,13 +24,9 @@ LPSTR __cdecl CRTDLL__mbsinc( LPCSTR str )
  */
 INT __cdecl CRTDLL__mbslen( LPCSTR str )
 {
-    INT len, total = 0;
-    while ((len = mblen( str, MB_LEN_MAX )) > 0)
-    {
-        str += len;
-        total++;
-    }
-    return total;
+    INT len;
+    for (len = 0; *str; len++, str++) if (IsDBCSLeadByte(str[0]) && str[1]) str++;
+    return len;
 }
 
 
@@ -49,8 +35,11 @@ INT __cdecl CRTDLL__mbslen( LPCSTR str )
  */
 INT __cdecl CRTDLL_mbtowc( WCHAR *dst, LPCSTR str, INT n )
 {
-    wchar_t res;
-    int ret = mbtowc( &res, str, n );
-    if (dst) *dst = (WCHAR)res;
-    return ret;
+    if (n <= 0) return 0;
+    if (!str) return 0;
+    if (!MultiByteToWideChar( CP_ACP, 0, str, n, dst, 1 )) return 0;
+    /* return the number of bytes from src that have been used */
+    if (!*str) return 0;
+    if (n >= 2 && IsDBCSLeadByte(*str) && str[1]) return 2;
+    return 1;
 }

dlls/crtdll/wcstring.c

@@ -11,6 +11,7 @@
 #include <string.h>
 
 #include "windef.h"
+#include "winnls.h"
 #include "wine/unicode.h"
 #include "crtdll.h"
 #include "debugtools.h"
@@ -113,17 +114,15 @@ LPWSTR __cdecl CRTDLL_wcspbrk( LPCWSTR str, LPCWSTR accept )
  */
 INT __cdecl CRTDLL_wctomb( LPSTR dst, WCHAR ch )
 {
-    return wctomb( dst, (wchar_t)ch );
+    return WideCharToMultiByte( CP_ACP, 0, ch, 1, dst, 6, NULL, NULL );
 }
 
-extern INT __cdecl NTDLL_iswctype( WCHAR wc, WCHAR wct ); /* FIXME */
-
 /*********************************************************************
  *           CRTDLL_iswalnum    (CRTDLL.405)
  */
 INT __cdecl CRTDLL_iswalnum( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0107 );
+    return get_char_typeW(wc) & (C1_ALPHA|C1_DIGIT|C1_LOWER|C1_UPPER);
 }
 
 /*********************************************************************
@@ -131,7 +130,7 @@ INT __cdecl CRTDLL_iswalnum( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswalpha( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0103 );
+    return get_char_typeW(wc) & (C1_ALPHA|C1_LOWER|C1_UPPER);
 }
 
 /*********************************************************************
@@ -139,7 +138,7 @@ INT __cdecl CRTDLL_iswalpha( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswcntrl( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0020 );
+    return get_char_typeW(wc) & C1_CNTRL;
 }
 
 /*********************************************************************
@@ -147,7 +146,7 @@ INT __cdecl CRTDLL_iswcntrl( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswdigit( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0004 );
+    return get_char_typeW(wc) & C1_DIGIT;
 }
 
 /*********************************************************************
@@ -155,7 +154,7 @@ INT __cdecl CRTDLL_iswdigit( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswgraph( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0117 );
+    return get_char_typeW(wc) & (C1_ALPHA|C1_PUNCT|C1_DIGIT|C1_LOWER|C1_UPPER);
 }
 
 /*********************************************************************
@@ -163,7 +162,7 @@ INT __cdecl CRTDLL_iswgraph( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswlower( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0002 );
+    return get_char_typeW(wc) & C1_LOWER;
 }
 
 /*********************************************************************
@@ -171,7 +170,7 @@ INT __cdecl CRTDLL_iswlower( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswprint( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0157 );
+    return get_char_typeW(wc) & (C1_ALPHA|C1_BLANK|C1_PUNCT|C1_DIGIT|C1_LOWER|C1_UPPER);
 }
 
 /*********************************************************************
@@ -179,7 +178,7 @@ INT __cdecl CRTDLL_iswprint( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswpunct( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0010 );
+    return get_char_typeW(wc) & C1_PUNCT;
 }
 
 /*********************************************************************
@@ -187,7 +186,7 @@ INT __cdecl CRTDLL_iswpunct( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswspace( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0008 );
+    return get_char_typeW(wc) & C1_SPACE;
 }
 
 /*********************************************************************
@@ -195,7 +194,7 @@ INT __cdecl CRTDLL_iswspace( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswupper( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0001 );
+    return get_char_typeW(wc) & C1_UPPER;
 }
 
 /*********************************************************************
@@ -203,5 +202,5 @@ INT __cdecl CRTDLL_iswupper( WCHAR wc )
  */
 INT __cdecl CRTDLL_iswxdigit( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0080 );
+    return get_char_typeW(wc) & C1_XDIGIT;
 }

dlls/ntdll/wcstring.c

@@ -9,9 +9,6 @@
 #include <ctype.h>
 #include <stdlib.h>
 #include <string.h>
-#ifdef HAVE_WCTYPE_H
-#include <wctype.h>
-#endif
 
 #include "windef.h"
 #include "winbase.h"
@@ -296,40 +293,7 @@ INT __cdecl NTDLL_wcstol(LPWSTR s,LPWSTR *end,INT base)
  */
 INT __cdecl NTDLL_iswctype( WCHAR wc, WCHAR wct )
 {
-    INT res = 0;
-
-#ifdef HAVE_WCTYPE_H
-#undef iswupper
-#undef iswlower
-#undef iswdigit
-#undef iswspace
-#undef iswpunct
-#undef iswcntrl
-#undef iswxdigit
-#undef iswalpha
-    if (wct & 0x0001) res |= iswupper(wc);
-    if (wct & 0x0002) res |= iswlower(wc);
-    if (wct & 0x0004) res |= iswdigit(wc);
-    if (wct & 0x0008) res |= iswspace(wc);
-    if (wct & 0x0010) res |= iswpunct(wc);
-    if (wct & 0x0020) res |= iswcntrl(wc);
-    if (wct & 0x0080) res |= iswxdigit(wc);
-    if (wct & 0x0100) res |= iswalpha(wc);
-#else
-    if (wct & 0x0001) res |= isupper(LOBYTE(wc));
-    if (wct & 0x0002) res |= islower(LOBYTE(wc));
-    if (wct & 0x0004) res |= isdigit(LOBYTE(wc));
-    if (wct & 0x0008) res |= isspace(LOBYTE(wc));
-    if (wct & 0x0010) res |= ispunct(LOBYTE(wc));
-    if (wct & 0x0020) res |= iscntrl(LOBYTE(wc));
-    if (wct & 0x0080) res |= isxdigit(LOBYTE(wc));
-    if (wct & 0x0100) res |= isalpha(LOBYTE(wc));
-#endif
-    if (wct & 0x0040)
-	FIXME(": iswctype(%04hx,_BLANK|...) requested\n",wc);
-    if (wct & 0x8000)
-	FIXME(": iswctype(%04hx,_LEADBYTE|...) requested\n",wc);
-    return res;
+    return (get_char_typeW(wc) & 0xfff) & wct;
 }
 
 
@@ -338,5 +302,5 @@ INT __cdecl NTDLL_iswctype( WCHAR wc, WCHAR wct )
  */
 INT __cdecl NTDLL_iswalpha( WCHAR wc )
 {
-    return NTDLL_iswctype( wc, 0x0100 );
+    return get_char_typeW(wc) & C1_ALPHA;
 }

dlls/user/lstr.c

@@ -320,6 +320,62 @@ BOOL WINAPI OemToCharW( LPCSTR s, LPWSTR d )
 
 
 /***********************************************************************
+ *           IsCharLowerA   (USER.436) (USER32.@)
+ * FIXME: handle current locale
+ */
+BOOL WINAPI IsCharLowerA(CHAR x)
+{
+    return islower(x);
+}
+
+
+/***********************************************************************
+ *           IsCharLowerW   (USER32.@)
+ */
+BOOL WINAPI IsCharLowerW(WCHAR x)
+{
+    return get_char_typeW(x) & C1_LOWER;
+}
+
+
+/***********************************************************************
+ *           IsCharUpperA   (USER.435) (USER32.337)
+ * FIXME: handle current locale
+ */
+BOOL WINAPI IsCharUpperA(CHAR x)
+{
+    return isupper(x);
+}
+
+
+/***********************************************************************
+ *           IsCharUpperW   (USER32.@)
+ */
+BOOL WINAPI IsCharUpperW(WCHAR x)
+{
+    return get_char_typeW(x) & C1_UPPER;
+}
+
+
+/***********************************************************************
+ *           IsCharAlphaNumericW   (USER32.@)
+ */
+BOOL WINAPI IsCharAlphaNumericW(WCHAR x)
+{
+    return get_char_typeW(x) & (C1_ALPHA|C1_DIGIT|C1_LOWER|C1_UPPER);
+}
+
+
+/***********************************************************************
+ *           IsCharAlphaW   (USER32.@)
+ */
+BOOL WINAPI IsCharAlphaW(WCHAR x)
+{
+    return get_char_typeW(x) & (C1_ALPHA|C1_LOWER|C1_UPPER);
+}
+
+
+/***********************************************************************
  *           FormatMessage16   (USER.606)
  */
 DWORD WINAPI FormatMessage16(

include/config.h.in

@@ -428,9 +428,6 @@
 /* Define if you have the <ucontext.h> header file.  */
 #undef HAVE_UCONTEXT_H
 
-/* Define if you have the <wctype.h> header file.  */
-#undef HAVE_WCTYPE_H
-
 /* Define if you have the curses library (-lcurses).  */
 #undef HAVE_LIBCURSES
 

include/winbase.h

@@ -1767,11 +1767,10 @@ INT       WINAPI GetProfileStringW(LPCWSTR,LPCWSTR,LPCWSTR,LPWSTR,UINT);
 VOID        WINAPI GetStartupInfoA(LPSTARTUPINFOA);
 VOID        WINAPI GetStartupInfoW(LPSTARTUPINFOW);
 #define     GetStartupInfo WINELIB_NAME_AW(GetStartupInfo)
-BOOL      WINAPI GetStringTypeA(LCID,DWORD,LPCSTR,INT,LPWORD);
-BOOL      WINAPI GetStringTypeW(DWORD,LPCWSTR,INT,LPWORD);
-#define     GetStringType WINELIB_NAME_AW(GetStringType)
-UINT      WINAPI GetSystemDirectoryA(LPSTR,UINT);
-UINT      WINAPI GetSystemDirectoryW(LPWSTR,UINT);
+BOOL        WINAPI GetStringTypeA(LCID,DWORD,LPCSTR,INT,LPWORD);
+BOOL        WINAPI GetStringTypeW(DWORD,LPCWSTR,INT,LPWORD);
+UINT        WINAPI GetSystemDirectoryA(LPSTR,UINT);
+UINT        WINAPI GetSystemDirectoryW(LPWSTR,UINT);
 #define     GetSystemDirectory WINELIB_NAME_AW(GetSystemDirectory)
 DWORD       WINAPI GetTickCount(void);
 UINT        WINAPI GetTempFileNameA(LPCSTR,LPCSTR,UINT,LPSTR);

include/wine/unicode.h

@@ -73,6 +73,14 @@ static inline WCHAR toupperW( WCHAR ch )
     return ch + casemap_upper[casemap_upper[ch >> 8] + (ch & 0xff)];
 }
 
+/* the character type contains the C1_* flags in the low 12 bits */
+/* and the C2_* type in the high 4 bits */
+static inline unsigned short get_char_typeW( WCHAR ch )
+{
+    extern unsigned short wctype_table[];
+    return wctype_table[wctype_table[ch >> 8] + (ch & 0xff)];
+}
+
 /* some useful string manipulation routines */
 
 static inline unsigned int strlenW( const WCHAR *str )

memory/codepage.c

@@ -360,3 +360,38 @@ INT WINAPI WideCharToMultiByte( UINT page, DWORD flags, LPCWSTR src, INT srclen,
     }
     return ret;
 }
+
+
+/******************************************************************************
+ *              GetStringTypeW   (KERNEL32)
+ *
+ */
+BOOL WINAPI GetStringTypeW( DWORD type, LPCWSTR src, INT count, LPWORD chartype )
+{
+    if (count == -1) count = strlenW(src) + 1;
+    switch(type)
+    {
+    case CT_CTYPE1:
+        while (count--) *chartype++ = get_char_typeW( *src++ ) & 0xfff;
+        break;
+    case CT_CTYPE2:
+        while (count--) *chartype++ = get_char_typeW( *src++ ) >> 12;
+        break;
+    case CT_CTYPE3:
+        FIXME("CT_CTYPE3 not supported.\n");
+    default:
+        SetLastError( ERROR_INVALID_PARAMETER );
+        return FALSE;
+    }
+    return TRUE;
+}
+
+
+/******************************************************************************
+ *              GetStringTypeExW   (KERNEL32)
+ */
+BOOL WINAPI GetStringTypeExW( LCID locale, DWORD type, LPCWSTR src, INT count, LPWORD chartype )
+{
+    /* locale is ignored for Unicode */
+    return GetStringTypeW( type, src, count, chartype );
+}

misc/lstr.c

@@ -13,15 +13,6 @@
 #include <string.h>
 #include <ctype.h>
 
-#ifdef HAVE_WCTYPE_H
-# include <wctype.h>
-#else
-# define iswalnum(c) isalnum(c)
-# define iswalpha(c) isalpha(c)
-# define iswupper(c) isupper(c)
-# define iswlower(c) islower(c)
-#endif  /* HAVE_WCTYPE_H */
-
 #include "windef.h"
 #include "winbase.h"
 #include "wingdi.h"
@@ -192,57 +183,3 @@ BOOL WINAPI IsCharAlphaNumericA(CHAR x)
 {
     return IsCharAlphaA(x) || isdigit(x) ;
 }
-
-/***********************************************************************
- *           IsCharAlphaNumericW   (USER32.333)
- * FIXME: handle current locale
- */
-BOOL WINAPI IsCharAlphaNumericW(WCHAR x)
-{
-    return iswalnum(x);
-}
-
-/***********************************************************************
- *           IsCharAlphaW   (USER32.334)
- * FIXME: handle current locale
- */
-BOOL WINAPI IsCharAlphaW(WCHAR x)
-{
-    return iswalpha(x);
-}
-
-/***********************************************************************
- *           IsCharLowerA   (USER.436) (USER32.335)
- * FIXME: handle current locale
- */
-BOOL WINAPI IsCharLowerA(CHAR x)
-{
-    return islower(x);
-}
-
-/***********************************************************************
- *           IsCharLowerW   (USER32.336)
- * FIXME: handle current locale
- */
-BOOL WINAPI IsCharLowerW(WCHAR x)
-{
-    return iswlower(x);
-}
-
-/***********************************************************************
- *           IsCharUpperA   (USER.435) (USER32.337)
- * FIXME: handle current locale
- */
-BOOL WINAPI IsCharUpperA(CHAR x)
-{
-    return isupper(x);
-}
-
-/***********************************************************************
- *           IsCharUpperW   (USER32.338)
- * FIXME: handle current locale
- */
-BOOL WINAPI IsCharUpperW(WCHAR x)
-{
-    return iswupper(x);
-}

ole/ole2nls.c

@@ -1223,60 +1223,6 @@ BOOL WINAPI GetStringTypeExA(LCID locale,DWORD dwInfoType,LPCSTR src,
 	}
 }
 
-/******************************************************************************
- *		GetStringTypeW	[KERNEL32.399]
- *
- * NOTES
- * Yes, this is missing LCID locale. MS fault.
- */
-BOOL WINAPI GetStringTypeW(DWORD dwInfoType,LPCWSTR src,INT cchSrc,
-                               LPWORD chartype)
-{
-	return GetStringTypeExW(0/*defaultlocale*/,dwInfoType,src,cchSrc,chartype);
-}
-
-/******************************************************************************
- *		GetStringTypeExW	[KERNEL32.398]
- *
- * FIXME: unicode chars are assumed chars
- */
-BOOL WINAPI GetStringTypeExW(LCID locale,DWORD dwInfoType,LPCWSTR src,
-                                 INT cchSrc,LPWORD chartype)
-{
-	int	i;
-	UINT	ch;
-
-	if (cchSrc==-1)
-	  cchSrc=lstrlenW(src)+1;
-	
-	switch (dwInfoType) {
-	case CT_CTYPE2:
-		FIXME("CT_CTYPE2 not supported.\n");
-		return FALSE;
-	case CT_CTYPE3:
-		FIXME("CT_CTYPE3 not supported.\n");
-		return FALSE;
-	default:break;
-	}
-	for (i=0;i<cchSrc;i++) {
-		chartype[i] = 0;
-		ch = src[i];
-		if ( ch >= (UINT)0x100 ) continue;
-		if (isdigit(ch)) chartype[i]|=C1_DIGIT;
-		if (isalpha(ch)) chartype[i]|=C1_ALPHA;
-		if (islower(ch)) chartype[i]|=C1_LOWER;
-		if (isupper(ch)) chartype[i]|=C1_UPPER;
-		if (isspace(ch)) chartype[i]|=C1_SPACE;
-		if (ispunct(ch)) chartype[i]|=C1_PUNCT;
-		if (iscntrl(ch)) chartype[i]|=C1_CNTRL;
-/* FIXME: isblank() is a GNU extension */
-/*		if (isblank(ch)) chartype[i]|=C1_BLANK; */
-                if ((ch == ' ') || (ch == '\t')) chartype[i]|=C1_BLANK;
-		/* C1_XDIGIT */
-	}
-	return TRUE;
-}
-
 /***********************************************************************
  *           VerLanguageNameA              [KERNEL32.709][VERSION.9]
  */

unicode/Makefile.in

@@ -72,6 +72,7 @@ C_SRCS = \
 	string.c \
 	utf8.c \
 	wctomb.c \
+	wctype.c \
 	$(CODEPAGES:%=c_%.c)
 
 all: libwine_unicode.$(LIBEXT)

unicode/cpmap.pl

@@ -76,11 +76,93 @@ $DEF_CHAR = ord '?';
     [ 28599, "ISO8859/8859-9.TXT",                "ISO 8859-9 Latin 5" ]
 );
 
+
+%ctype =
+(
+    "upper"  => 0x0001,
+    "lower"  => 0x0002,
+    "digit"  => 0x0004,
+    "space"  => 0x0008,
+    "punct"  => 0x0010,
+    "cntrl"  => 0x0020,
+    "blank"  => 0x0040,
+    "xdigit" => 0x0080,
+    "alpha"  => 0x0100
+);
+
+%categories =
+(
+    "Lu" => $ctype{"alpha"}|$ctype{"upper"}, # Letter, Uppercase
+    "Ll" => $ctype{"alpha"}|$ctype{"lower"}, # Letter, Lowercase
+    "Lt" => $ctype{"alpha"},    # Letter, Titlecase
+    "Mn" => $ctype{"punct"},    # Mark, Non-Spacing
+    "Mc" => $ctype{"punct"},    # Mark, Spacing Combining
+    "Me" => $ctype{"punct"},    # Mark, Enclosing
+    "Nd" => $ctype{"digit"},    # Number, Decimal Digit
+    "Nl" => $ctype{"punct"},    # Number, Letter
+    "No" => $ctype{"punct"},    # Number, Other
+    "Zs" => $ctype{"space"},    # Separator, Space
+    "Zl" => 0,                  # Separator, Line
+    "Zp" => 0,                  # Separator, Paragraph
+    "Cc" => $ctype{"cntrl"},    # Other, Control
+    "Cf" => 0,                  # Other, Format
+    "Cs" => 0,                  # Other, Surrogate
+    "Co" => 0,                  # Other, Private Use
+    "Cn" => 0,                  # Other, Not Assigned
+    "Lm" => $ctype{"punct"},    # Letter, Modifier
+    "Lo" => $ctype{"alpha"},    # Letter, Other
+    "Pc" => $ctype{"punct"},    # Punctuation, Connector
+    "Pd" => $ctype{"punct"},    # Punctuation, Dash
+    "Ps" => $ctype{"punct"},    # Punctuation, Open
+    "Pe" => $ctype{"punct"},    # Punctuation, Close
+    "Pi" => $ctype{"punct"},    # Punctuation, Initial quote
+    "Pf" => $ctype{"punct"},    # Punctuation, Final quote
+    "Po" => $ctype{"punct"},    # Punctuation, Other
+    "Sm" => $ctype{"punct"},    # Symbol, Math
+    "Sc" => $ctype{"punct"},    # Symbol, Currency
+    "Sk" => $ctype{"punct"},    # Symbol, Modifier
+    "So" => $ctype{"punct"}     # Symbol, Other 
+);
+
+# a few characters need additional categories that cannot be determined automatically
+%special_categories =
+(
+    "xdigit" => [ ord('0')..ord('9'),ord('A')..ord('F'),ord('a')..ord('f'),
+                  0xff10..0xff19, 0xff21..0xff26, 0xff41..0xff46 ],
+    "space"  => [ 0x09..0x0d, 0xfeff ],
+    "blank"  => [ 0x09, 0x20, 0xa0, 0xfeff ]
+);
+
+%directions =
+(
+    "L"   => 1,    # Left-to-Right
+    "LRE" => 11,   # Left-to-Right Embedding
+    "LRO" => 11,   # Left-to-Right Override
+    "R"   => 2,    # Right-to-Left
+    "AL"  => 2,    # Right-to-Left Arabic
+    "RLE" => 11,   # Right-to-Left Embedding
+    "RLO" => 11,   # Right-to-Left Override
+    "PDF" => 11,   # Pop Directional Format
+    "EN"  => 3,    # European Number
+    "ES"  => 4,    # European Number Separator
+    "ET"  => 5,    # European Number Terminator
+    "AN"  => 6,    # Arabic Number
+    "CS"  => 7,    # Common Number Separator
+    "NSM" => 0,    # Non-Spacing Mark
+    "BN"  => 0,    # Boundary Neutral
+    "B"   => 8,    # Paragraph Separator
+    "S"   => 9,    # Segment Separator
+    "WS"  => 10,   # Whitespace
+    "ON"  => 11    # Other Neutrals
+);
+
+
 ################################################################
 # main routine
 
 READ_DEFAULTS();
 DUMP_CASE_MAPPINGS();
+DUMP_CTYPE_TABLES();
 
 foreach $file (@allfiles) { HANDLE_FILE( @$file ); }
 
@@ -97,6 +179,8 @@ sub READ_DEFAULTS
     @unicode_aliases = ();
     @tolower_table = ();
     @toupper_table = ();
+    @category_table = ();
+    @direction_table = ();
 
     # first setup a few default mappings
 
@@ -135,12 +219,43 @@ sub READ_DEFAULTS
 	 $decomp, $dec, $dig, $num, $mirror, 
 	 $oldname, $comment, $upper, $lower, $title) = split /;/;
 
-        if ($lower ne "") { $tolower_table[hex $code] = hex $lower; }
-        if ($upper ne "") { $toupper_table[hex $code] = hex $upper; }
+        my $src = hex $code;
 
-        next if $decomp eq "";  # no decomposition, skip it
+        die "unknown category $cat" unless defined $categories{$cat};
+        die "unknown directionality $bidi" unless defined $directions{$bidi};
+
+        $uniname[$src] = $name;
+        $category_table[$src] = $categories{$cat};
+        $direction_table[$src] = $directions{$bidi};
+
+        if ($lower ne "")
+        {
+            $tolower_table[$src] = hex $lower;
+            $category_table[$src] |= $ctype{"upper"}|$ctype{"alpha"};
+        }
+        if ($upper ne "")
+        {
+            $toupper_table[$src] = hex $upper;
+            $category_table[$src] |= $ctype{"lower"}|$ctype{"alpha"};
+        }
+        if ($dec ne "")
+        {
+            $category_table[$src] |= $ctype{"digit"};
+        }
 
-        $src = hex $code;
+        # copy the category and direction for everything between First/Last pairs
+        if ($name =~ /, First>/) { $start = $src; }
+        if ($name =~ /, Last>/)
+        {
+            while ($start < $src)
+            {
+                $category_table[$start] = $category_table[$src];
+                $direction_table[$start] = $direction_table[$src];
+                $start++;
+            }
+        }
+        
+        next if $decomp eq "";  # no decomposition, skip it
 
         if ($decomp =~ /^<([a-zA-Z]+)>\s+([0-9a-fA-F]+)$/)
         {
@@ -165,6 +280,7 @@ sub READ_DEFAULTS
         {
             # decomposition contains only char values without prefix -> use first char
             $dst = hex $1;
+            $category_table[$src] |= $category_table[$dst];
         }
         else
         {
@@ -181,6 +297,14 @@ sub READ_DEFAULTS
         }
         $unicode_defaults[$src] = $dst;
     }
+
+    # patch the category of some special characters
+
+    foreach $cat (keys %special_categories)
+    {
+        my $flag = $ctype{$cat};
+        foreach $i (@{$special_categories{$cat}}) { $category_table[$i] |= $flag; }
+    }
 }
 
 
@@ -567,6 +691,49 @@ sub DUMP_CASE_TABLE
 
 
 ################################################################
+# dump the ctype tables
+sub DUMP_CTYPE_TABLES
+{
+    open OUTPUT,">wctype.c" or die "Cannot create casemap.c";
+    printf "Building wctype.c\n";
+    printf OUTPUT "/* Unicode ctype tables */\n";
+    printf OUTPUT "/* Automatically generated; DO NOT EDIT!! */\n\n";
+    printf OUTPUT "#include \"wine/unicode.h\"\n\n";
+
+    my $i;
+    my @array = (0) x 256;
+
+    # add the direction in the high 4 bits of the category
+    for ($i = 0; $i < 65536; $i++)
+    {
+        $category_table[$i] |= $direction_table[$i] << 12;
+    }
+
+    # try to merge table rows
+    for ($row = 0; $row < 256; $row++)
+    {
+        my $rowtxt = sprintf "%04x" x 256, @category_table[($row<<8)..($row<<8)+255];
+        if (defined($sequences{$rowtxt}))
+        {
+            # reuse an existing row
+            $array[$row] = $sequences{$rowtxt};
+        }
+        else
+        {
+            # create a new row
+            $sequences{$rowtxt} = $array[$row] = $#array + 1;
+            push @array, @category_table[($row<<8)..($row<<8)+255];
+        }
+    }
+
+    printf OUTPUT "const unsigned short wctype_table[%d] =\n{\n", $#array+1;
+    printf OUTPUT "    /* offsets */\n%s,\n", DUMP_ARRAY( "0x%04x", 0, @array[0..255] );
+    printf OUTPUT "    /* values */\n%s\n};\n", DUMP_ARRAY( "0x%04x", 0, @array[256..$#array] );
+
+    close OUTPUT;
+}
+
+################################################################
 # read an input file and generate the corresponding .c file
 sub HANDLE_FILE
 {