windows-server-2003/windows/richedit/re30/w32sys.cpp

/*
 *	@doc INTERNAL
 *
 *	@module	w32sys.cpp - thin layer over Win32 services
 *	
 *	History: <nl>
 *		1/22/97 joseogl Created
 *
 *	Copyright (c) 1995-1998 Microsoft Corporation. All rights reserved.
 */

// This prevents the "W32->" prefix from being prepended to our identifiers.

#define W32SYS_CPP

#include "_common.h"
#include "_host.h"
#include "_font.h"
#include "_edit.h"
#include <malloc.h>

// Include the appropriate implementation.
#if defined(PEGASUS)
#include "w32wince.cpp"
#else
#include "w32win32.cpp"
#endif

ASSERTDATA

/*
 *	@struct CPGCHAR |
 *		Locally used variable that contains code-page and char-set info
 */
typedef struct _cpgcharset
{
	INT			nCodePage;				// @field Code page
	BYTE		bCharSet;				// @field Character set
	DWORD		dwFontSig;				// @field Font signature bits
} CPGCHAR;

static const CPGCHAR rgCpgCharSet[NCHARSETS] =
{
	{1252,	ANSI_CHARSET,			fLATIN1>>8}, // fLATIN1 has 3 bits
	{0,		DEFAULT_CHARSET,		0x00000000}, // Not reliably implemented...
	{CP_SYMBOL,SYMBOL_CHARSET,		0x00004000}, // No trans, except WORD -> BYTE
	{437,	PC437_CHARSET,			0x00000000}, // United States IBM
	{850,	OEM_CHARSET,			0x00000400}, // IBM Multilingual
	{1250,	EASTEUROPE_CHARSET,		0x00000002}, // Eastern Europe
	{1255,	HEBREW_CHARSET,			0x00000020}, // Hebrew
	{1256,  ARABIC_CHARSET,			0x00000040}, // Arabic
	{932,	SHIFTJIS_CHARSET,		0x00020000}, // Japanese
	{1251,	RUSSIAN_CHARSET,		0x00000004}, // Russian
	{936,	GB2312_CHARSET,			0x00040000}, // PRC
	{949,	HANGEUL_CHARSET,		0x00080000}, // Hangul
	{1361,	JOHAB_CHARSET,			0x00000000}, // JOHAB
	{950,	CHINESEBIG5_CHARSET,	0x00100000}, // Chinese
	{1253,	GREEK_CHARSET,			0x00000008}, // Greek
	{1254,	TURKISH_CHARSET,		0x00000010}, // Turkish
	{1257,	BALTIC_CHARSET,			0x00000080}, // Estonia, Lithuania, Latvia
	{1258,	VIETNAMESE_CHARSET,		0x00000100}, // Vietnamese
	{874,	THAI_CHARSET,			0x00010000}, // Thai
	{CP_DEVANAGARI,DEVANAGARI_CHARSET,0x00200000}, // Devanagari
	{CP_TAMIL, TAMIL_CHARSET,		0x00400000}, // Tamil
	{CP_GEORGIAN, GEORGIAN_CHARSET,	0x00800000}, // Georgian
	{CP_ARMENIAN, ARMENIAN_CHARSET,	0x00000200}, // Armenian
	{10000,	MAC_CHARSET,			0x00000000}	 // Most popular Mac (English, etc.)
};

#define cCpgCharSet ARRAY_SIZE(rgCpgCharSet)

#define	LANG_TAIWAN	 MAKELANGID(LANG_CHINESE, SUBLANG_CHINESE_TRADITIONAL)

const WORD CodePageTable[] = {
/* CodePage		  PLID	primary language
   ------------------------------------- */
	   0,		// 00 -	undefined
	1256,		// 01 - Arabic
	1251,		// 02 - Bulgarian
	1252,		// 03 - Catalan
	 950,		// 04 - Taiwan (Hong Kong, PRC, and Singapore are 936)
	1250,		// 05 - Czech
	1252,		// 06 - Danish
	1252,		// 07 - German
	1253,		// 08 - Greek
	1252,		// 09 - English
	1252,		// 0a - Spanish
	1252,		// 0b - Finnish
	1252,		// 0c - French
	1255,		// 0d - Hebrew
	1250,		// 0e - Hungarian
	1252,		// 0f - Icelandic
	1252,		// 10 - Italian
	 932,		// 11 - Japan
	 949,		// 12 - Korea
	1252,		// 13 - Dutch
	1252,		// 14 - Norwegian
	1250,		// 15 - Polish
	1252,		// 16 - Portuguese
	   0,		// 17 -	Rhaeto-Romanic
	1250,		// 18 - Romanian
	1251,		// 19 - Russian
	1250,		// 1a -	Croatian
	1250,		// 1b - Slovak
	1250,		// 1c -	Albanian
	1252,		// 1d - Swedish
	 874,		// 1e - Thai
	1254,		// 1f - Turkish
	1256,		// 20 -	Urdu
	1252,		// 21 - Indonesian
	1251,		// 22 - Ukranian
	1251,		// 23 - Byelorussian
	1250,		// 24 -	Slovenian
	1257,		// 25 - Estonia
	1257,		// 26 - Latvian
	1257,		// 27 - Lithuanian
	   0,		// 28 -	Tajik - Tajikistan (undefined)
	1256,		// 29 - Farsi
	1258,		// 2a -	Vietnanese
	CP_ARMENIAN,// 2b -	Armenian (Unicode only)
	1254,		// 2c -	Azeri (Latin, can be Cyrillic...)
	1252,		// 2d - Basque
	   0,		// 2e - Sorbian
	1251,		// 2f - FYRO Macedonian
	1252,		// 30 - Sutu
	1252,		// 31 - Tsonga
	1252,		// 32 - Tswana
	1252,		// 33 - Venda
	1252,		// 34 - Xhosa
	1252,		// 35 - Zulu
	1252,		// 36 - Africaans
	CP_GEORGIAN,// 37 - Georgian (Unicode only)
	1252,		// 38 - Faerose
	CP_DEVANAGARI,// 39 - Hindi (Indic)
	1252,		// 3a - Maltese
	1252,		// 3b - Sami
	1252,		// 3c - Gaelic
	1255,		// 3d - Yiddish
	1252,		// 3e - Malaysian
	1251,		// 3f - Kazakh
	1252,		// 40 -	Kirghiz
	1252,		// 41 - Swahili
	1252,		// 42 - Turkmen
	1254,		// 43 - Uzbek (Latin, can be Cyrillic...)
	1251		// 44 - Tatar
				// 45 - Bengali (Indic)
				// 46 - Punjabi(Gurmukhi) (Indic)
				// 47 - Gujarati (Indic)
				// 48 - Oriya (Indic)
				// 49 - Tamil (Indic)
				// 4a - Telugu (Indic)
				// 4b - Kannada (Indic)
				// 4c - Malayalam (Indic)
				// 4d - Assamese (Indic)
				// 4e - Marathi (Indic)
				// 4f - Sanskrit (Indic)
				// 50*  Mongolian - Mongolia
				// 51*  Tibetan - Tibet
				// 52*  Welsh - Wales
				// 53*  Khmer - Cambodia
				// 54*  Lao - Lao
				// 55*  Burmese - Mayanmar
				// 56*  Gallego - Portugal
				// 57 - Konkani (Indic)
				// 58*  Manipuri (Indic)
				// 59*  Sindhi (Indic)
				// 5a
				// 5b
				// 5c
				// 5d
				// 5e
				// 5f
				// 60*  Kashmiri
				// 61*  Nepali - Nepal
				// 62*  Frisian - Netherlands
				// 63*  Pashto - Afghanistan
};

#define nCodePageTable	ARRAY_SIZE(CodePageTable)
#define lidAzeriCyrillic 0x82C
#define lidSerbianCyrillic 0xC1A
#define lidUzbekCyrillic 0x843

// Our interface pointer
CW32System *W32;

CW32System::CW32System( )
{
	if(GetVersion(&_dwPlatformId, &_dwMajorVersion, &_dwMinorVersion))
	{
		_fHaveAIMM = FALSE;
		_fHaveIMMEShare = FALSE;
		_fHaveIMMProcs = FALSE;
		_pIMEShare = NULL;
		_icr3DDarkShadow = COLOR_WINDOWFRAME;
		if(_dwMajorVersion >= VERS4)
			_icr3DDarkShadow = COLOR_3DDKSHADOW;
	}

	_syslcid = GetSystemDefaultLCID();
	_ACP = ::GetACP();

    // BUG FIX #6089
    // we need this for backward compatibility of mouse wheel
    _MSMouseRoller = RegisterWindowMessageA(MSH_MOUSEWHEEL);

	// Register private messages for MSIME98	
	_MSIMEMouseMsg = RegisterWindowMessageA("MSIMEMouseOperation");
	_MSIMEDocFeedMsg = RegisterWindowMessageA("MSIMEDocumentFeed");	
	_MSIMEQueryPositionMsg = RegisterWindowMessageA("MSIMEQueryPosition");	
	_MSIMEServiceMsg = RegisterWindowMessageA("MSIMEService");

	// get MSIME Reconvert private messages unless we are running in NT5
	if (_dwPlatformId == VER_PLATFORM_WIN32_WINDOWS ||
		(_dwPlatformId == VER_PLATFORM_WIN32_NT && _dwMajorVersion <= 4))
	{
		_MSIMEReconvertMsg = RegisterWindowMessageA("MSIMEReconvert");
		_MSIMEReconvertRequestMsg = RegisterWindowMessageA("MSIMEReconvertRequest");

	}
	else
	{
		_MSIMEReconvertMsg = 0;			// private message for reconversion
		_MSIMEReconvertRequestMsg = 0;	// private message for reconversion request
	}
}

CW32System::~CW32System()
{
	FreeIME();	
	FreeOle();
	if (_hdcScreen)
		DeleteDC(_hdcScreen);
}

///////////////////////////////  Memory  and CRT utility functions  /////////////////////////////////
extern "C" {

#ifdef NOCRTOBJS

// Havinf these functions defined here helps eliminate the dependency on the CRT
// Some function definitions copied from CRT sources.
// Typically, it is better to get the objs for these objects from the CRT
// without dragging in the whole thing.

/***
*int memcmp(buf1, buf2, count) - compare memory for lexical order
*
*Purpose:
*       Compares count bytes of memory starting at buf1 and buf2
*       and find if equal or which one is first in lexical order.
*
*Entry:
*       void *buf1, *buf2 - pointers to memory sections to compare
*       size_t count - length of sections to compare
*
*Exit:
*       returns < 0 if buf1 < buf2
*       returns  0  if buf1 == buf2
*       returns > 0 if buf1 > buf2
*
*Exceptions:
*
*******************************************************************************/

int __cdecl memcmp (
        const void * buf1,
        const void * buf2,
        size_t count
        )
{
        if (!count)
                return(0);

        while ( --count && *(char *)buf1 == *(char *)buf2 ) {
                buf1 = (char *)buf1 + 1;
                buf2 = (char *)buf2 + 1;
        }

        return( *((unsigned char *)buf1) - *((unsigned char *)buf2) );
}

/***
*char *memset(dst, val, count) - sets "count" bytes at "dst" to "val"
*
*Purpose:
*       Sets the first "count" bytes of the memory starting
*       at "dst" to the character value "val".
*
*Entry:
*       void *dst - pointer to memory to fill with val
*       int val   - value to put in dst bytes
*       size_t count - number of bytes of dst to fill
*
*Exit:
*       returns dst, with filled bytes
*
*Exceptions:
*
*******************************************************************************/

void * __cdecl memset (
        void *dst,
        int val,
        size_t count
        )
{
        void *start = dst;

        while (count--) {
                *(char *)dst = (char)val;
                dst = (char *)dst + 1;
        }

        return(start);
}

/***
*memcpy - Copy source buffer to destination buffer
*
*Purpose:
*       memcpy() copies a source memory buffer to a destination memory buffer.
*       This routine does NOT recognize overlapping buffers, and thus can lead
*       to propogation.
*
*       For cases where propogation must be avoided, memmove() must be used.
*
*Entry:
*       void *dst = pointer to destination buffer
*       const void *src = pointer to source buffer
*       size_t count = number of bytes to copy
*
*Exit:
*       Returns a pointer to the destination buffer
*
*Exceptions:
*******************************************************************************/

void * __cdecl memcpy (
        void * dst,
        const void * src,
        size_t count
        )
{
        void * ret = dst;

        /*
         * copy from lower addresses to higher addresses
         */
        while (count--) {
                *(char *)dst = *(char *)src;
                dst = (char *)dst + 1;
                src = (char *)src + 1;
        }

        return(ret);
}

void * __cdecl memmove(void *dst, const void *src, size_t count)
{
	void * ret = dst;

	if (dst <= src || (char *)dst >= ((char *)src + count)) {
		/*
         * Non-Overlapping Buffers
         * copy from lower addresses to higher addresses
         */
         while (count--) {
			*(char *)dst = *(char *)src;
            dst = (char *)dst + 1;
            src = (char *)src + 1;
         }
	}
    else
	{
		/*
         * Overlapping Buffers
         * copy from higher addresses to lower addresses
         */
        dst = (char *)dst + count - 1;
        src = (char *)src + count - 1;

        while (count--) {
			*(char *)dst = *(char *)src;
            dst = (char *)dst - 1;
            src = (char *)src - 1;
        }
	}

	return(ret);
}

/***
*strlen - return the length of a null-terminated string
*
*Purpose:
*       Finds the length in bytes of the given string, not including
*       the final null character.
*
*Entry:
*       const char * str - string whose length is to be computed
*
*Exit:
*       length of the string "str", exclusive of the final null byte
*
*Exceptions:
*
*******************************************************************************/

size_t __cdecl strlen (
        const char * str
        )
{
        const char *eos = str;

        while( *eos++ ) ;

        return( (int)(eos - str - 1) );
}

#endif

#ifdef DEBUG

// These functions are only used for RTF logging

/***
*strcmp - compare two strings, returning less than, equal to, or greater than
*
*Purpose:
*       STRCMP compares two strings and returns an integer
*       to indicate whether the first is less than the second, the two are
*       equal, or whether the first is greater than the second.
*
*       Comparison is done byte by byte on an UNSIGNED basis, which is to
*       say that Null (0) is less than any other character (1-255).
*
*Entry:
*       const char * src - string for left-hand side of comparison
*       const char * dst - string for right-hand side of comparison
*
*Exit:
*       returns -1 if src <  dst
*       returns  0 if src == dst
*       returns +1 if src >  dst
*
*Exceptions:
*
*******************************************************************************/

int __cdecl strcmp (
        const char * src,
        const char * dst
        )
{
        int ret = 0 ;

        while( ! (ret = *(unsigned char *)src - *(unsigned char *)dst) && *dst)
                ++src, ++dst;

        if ( ret < 0 )
                ret = -1 ;
        else if ( ret > 0 )
                ret = 1 ;

        return( ret );
}

/***
*char *strcat(dst, src) - concatenate (append) one string to another
*
*Purpose:
*       Concatenates src onto the end of dest.  Assumes enough
*       space in dest.
*
*Entry:
*       char *dst - string to which "src" is to be appended
*       const char *src - string to be appended to the end of "dst"
*
*Exit:
*       The address of "dst"
*
*Exceptions:
*
*******************************************************************************/

char * __cdecl strcat (
        char * dst,
        const char * src
        )
{
        char * cp = dst;

        while( *cp )
                cp++;                   /* find end of dst */

        while( *cp++ = *src++ ) ;       /* Copy src to end of dst */

        return( dst );                  /* return dst */

}

#endif

// This function in the runtime traps virtial base calls
int __cdecl _purecall()
{
	MessageBox(NULL,TEXT("Fatal Error : Vrtual base call in RichEdit"),NULL, MB_OK);
	PostQuitMessage (0);
	return 0;
}

// To avoid brionging in floating point lib
extern int _fltused = 1;

} // end of extern "C" block

size_t CW32System::wcslen(const wchar_t *wcs)
{
        const wchar_t *eos = wcs;

        while( *eos++ ) ;

        return( (size_t)(eos - wcs - 1) );
}

wchar_t * CW32System::wcscpy(wchar_t * dst, const wchar_t * src)
{
        wchar_t * cp = dst;

        while( *cp++ = *src++ )
                ;               /* Copy src over dst */

        return( dst );
}

int CW32System::wcscmp(const wchar_t * src, const wchar_t * dst)
{
	int ret = 0;

	while( ! (ret = (int)(*src - *dst)) && *dst)
		++src, ++dst;

	if ( ret < 0 )
		ret = -1 ;
	else if ( ret > 0 )
		ret = 1 ;

	return( ret );
}

int CW32System::wcsicmp(const wchar_t * src, const wchar_t * dst)
{
	int ret = 0;
	wchar_t s,d;
	
	do
	{
		s = ((*src <= L'Z') && (*dst >= L'A'))
			? *src - L'A' + L'a'
			: *src;
		d = ((*dst <= L'Z') && (*dst >= L'A'))
			? *dst - L'A' + L'a'
			: *dst;
		src++;
		dst++;
	} while (!(ret = (int)(s - d)) && d);

	if ( ret < 0 )
		ret = -1 ;
	else if ( ret > 0 )
		ret = 1 ;

	return( ret );
}

wchar_t * CW32System::wcsncpy (wchar_t * dest, const wchar_t * source, size_t count)
{
	wchar_t *start = dest;
	
	while (count && (*dest++ = *source++))	  /* copy string */
		count--;
	
	if (count)								/* pad out with zeroes */
		while (--count)
			*dest++ = L'\0';
		
	return(start);
}

int CW32System::wcsnicmp (const wchar_t * first, const wchar_t * last, size_t count)
{
	wchar_t f,l;
	int result = 0;
	
	if ( count ) {
		do {
			f = ((*first <= L'Z') && (*first >= L'A'))
				? *first - L'A' + L'a'
				: *first;
			l = ((*last <= L'Z') && (*last >= L'A'))
				? *last - L'A' + L'a'
				: *last;
			first++;
			last++;
		} while ( (--count) && f && (f == l) );
		result = (int)(f - l);
	}
	return result;
}

unsigned long CW32System::strtoul(const char *nptr)
{
	const char *p;
	char c;
	unsigned long number;
	unsigned digval;
	unsigned long maxval;
	
	p = nptr;                       /* p is our scanning pointer */
	number = 0;                     /* start with zero */
	
	c = *p++;                       /* read char */
	while ( c == ' ' || c == '\t' )
		c = *p++;               /* skip whitespace */
	
	if (c == '-') {
		return 0;
	}
		
	/* if our number exceeds this, we will overflow on multiply */
	maxval = ULONG_MAX / 10;
	
	for (;;) {      /* exit in middle of loop */
		/* convert c to value */
		digval = (unsigned char) c;
		if ( digval >= '0' && digval <= '9' )
			digval = c - '0';
		else
			return number;
		
		/* we now need to compute number = number * base + digval,
		but we need to know if overflow occured.  This requires
		a tricky pre-check. */
		
		if (number < maxval || (number == maxval &&
			(unsigned long)digval <= ULONG_MAX % 10)) {
			/* we won't overflow, go ahead and multiply */
			number = number * 10 + digval;
		}
		else
			return 0;
		
		c = *p++;               /* read next digit */
	}
}


// CW32System static members
BOOL	CW32System::_fLRMorRLM;
BOOL	CW32System::_fHaveIMMProcs;
BOOL	CW32System::_fHaveIMMEShare;
BOOL	CW32System::_fHaveAIMM;
UINT	CW32System::_fRegisteredXBox;
DWORD	CW32System::_dwPlatformId;
LCID	CW32System::_syslcid;
DWORD	CW32System::_dwMajorVersion;
DWORD	CW32System::_dwMinorVersion;
INT		CW32System::_icr3DDarkShadow;
UINT	CW32System::_MSIMEMouseMsg;	
UINT	CW32System::_MSIMEReconvertMsg;
UINT	CW32System::_MSIMEReconvertRequestMsg;	
UINT	CW32System::_MSIMEDocFeedMsg;
UINT	CW32System::_MSIMEQueryPositionMsg;
UINT	CW32System::_MSIMEServiceMsg;
UINT	CW32System::_MSMouseRoller;
HDC		CW32System::_hdcScreen;
CIMEShare* CW32System::_pIMEShare;

// CW32System static system parameter members
BOOL	CW32System::_fSysParamsOk;
BOOL 	CW32System::_fUsePalette;
INT 	CW32System::_xWidthSys;
INT 	CW32System::_yHeightSys;
INT		CW32System::_ySysFontLeading;
LONG 	CW32System::_xPerInchScreenDC;
LONG 	CW32System::_yPerInchScreenDC;
INT		CW32System::_cxBorder;
INT		CW32System::_cyBorder;
INT		CW32System::_cxVScroll;
INT		CW32System::_cyHScroll;
LONG 	CW32System::_dxSelBar;
INT		CW32System::_cxDoubleClk;
INT		CW32System::_cyDoubleClk;	
INT		CW32System::_DCT;
WORD	CW32System::_nScrollInset;
WORD	CW32System::_nScrollDelay;
WORD	CW32System::_nScrollInterval;
WORD	CW32System::_nScrollHAmount;
WORD	CW32System::_nScrollVAmount;
WORD	CW32System::_nDragDelay;
WORD	CW32System::_nDragMinDist;
WORD	CW32System::_wDeadKey;
WORD	CW32System::_wKeyboardFlags;
WORD	CW32System::_wNumKeyPad;
WORD	CW32System::_fFEFontInfo;
BYTE	CW32System::_bDigitSubstMode;
BYTE	CW32System::_bSysCharSet;
HCURSOR CW32System::_hcurSizeNS;
HCURSOR CW32System::_hcurSizeWE;
HCURSOR CW32System::_hcurSizeNWSE;
HCURSOR CW32System::_hcurSizeNESW;
LONG	CW32System::_cLineScroll;
HFONT	CW32System::_hSystemFont;
HKL		CW32System::_hklCurrent;
HKL		CW32System::_hkl[NCHARSETS];
INT		CW32System::_sysiniflags;
UINT	CW32System::_ACP;

DWORD	CW32System::_cRefs;


/*
 *  CW32System::MbcsFromUnicode(pstr, cch, pwstr, cwch, codepage, flags)
 *
 *  @mfunc
 *		Converts a string to MBCS from Unicode. If cwch equals -1, the string
 *		is assumed to be NULL terminated.  -1 is supplied as a default argument.
 *
 *	@rdesc
 *		If [pstr] is NULL or [cch] is 0, 0 is returned.  Otherwise, the number
 *		of characters converted, including the terminating NULL, is returned
 *		(note that converting the empty string will return 1).  If the
 *		conversion fails, 0 is returned.
 *
 *	@devnote
 *		Modifies pstr
 */
int CW32System::MbcsFromUnicode(
	LPSTR	pstr,		//@parm Buffer for MBCS string
	int		cch,		//@parm Size of MBCS buffer, incl space for NULL terminator
	LPCWSTR pwstr,		//@parm Unicode string to convert
	int		cwch,		//@parm # chars in Unicode string, incl NULL terminator
	UINT	codepage,	//@parm Code page to use (CP_ACP is default)
	UN_FLAGS flags)		//@parm Indicates if WCH_EMBEDDING should be handled specially
{
	TRACEBEGIN(TRCSUBSYSWRAP, TRCSCOPEINTERN, "CW32System::MbcsFromUnicode");

	LONG			i;
	LPWSTR			pwstrtemp;
	CTempWcharBuf	twcb;

    Assert(cch >= 0 && pwstr && (cwch == -1 || cwch > 0));

    if(!pstr || !cch)
        return 0;

	// If we have to convert WCH_EMBEDDINGs, scan through and turn
	// them into spaces.  This is necessary for richedit1.0 compatibity,
	// as WideCharToMultiByte will turn WCH_EMBEDDING into a '?'
	if(flags == UN_CONVERT_WCH_EMBEDDING)
	{
		if(cwch == -1)
			cwch = wcslen(pwstr) + 1;

		pwstrtemp = twcb.GetBuf(cwch);
		if(pwstrtemp)
		{
			for(i = 0; i < cwch; i++)
			{
				pwstrtemp[i] = pwstr[i];

				if(pwstr[i] == WCH_EMBEDDING)
					pwstrtemp[i] = L' ';
			}
			pwstr = pwstrtemp;
		}
	}
    return WCTMB(codepage, 0, pwstr, cwch, pstr, cch, NULL, NULL, NULL);
}

/*
 *  CW32System::UnicodeFromMbcs(pwstr, cwch, pstr, cch,	uiCodePage)
 *
 *	@mfunc
 *		Converts a string to Unicode from MBCS.  If cch equals -1, the string
 *		is assumed to be NULL terminated.  -1 is supplied as a default
 *		argument.
 *
 *	@rdesc
 *		If [pwstr] is NULL or [cwch] is 0, 0 is returned.  Otherwise,
 *		the number of characters converted, including the terminating
 *		NULL, is returned (note that converting the empty string will
 *		return 1).  If the conversion fails, 0 is returned.
 *
 *	@devnote
 *		Modifies:   [pwstr]
 */
int CW32System::UnicodeFromMbcs(
	LPWSTR	pwstr,		//@parm Buffer for Unicode string
	int		cwch,		//@parm Size of Unicode buffer, incl space for NULL terminator
	LPCSTR	pstr,		//@parm MBCS string to convert
	int		cch,		//@parm # chars in MBCS string, incl NULL terminator
	UINT	uiCodePage)	//@parm Code page to use (CP_ACP is default)
{
	TRACEBEGIN(TRCSUBSYSWRAP, TRCSCOPEINTERN, "CW32System::UnicodeFromMbcs");

    Assert(pstr && cwch >= 0 && (cch == -1 || cch >= 0));

    if(!pwstr || !cwch)
        return 0;

	if(cch >= 3 && IsUTF8BOM((BYTE *)pstr))
	{
		uiCodePage = CP_UTF8;				// UTF-8 BOM file
		cch -= 3;							// Eat the BOM
		pstr += 3;
	}
    return MBTWC(uiCodePage, 0, pstr, cch, pwstr, cwch, NULL);
}

/*
 *	CW32System::TextHGlobalAtoW (hglobalA)
 *
 *	@func
 *		translates a unicode string contained in an hglobal and
 *		wraps the ansi version in another hglobal
 *
 *	@devnote
 *		does *not* free the incoming hglobal
 */
HGLOBAL	CW32System::TextHGlobalAtoW(HGLOBAL hglobalA)
{
	TRACEBEGIN(TRCSUBSYSWRAP, TRCSCOPEINTERN, "CW32System::TextHGlobalAtoW");

	if(!hglobalA)
		return NULL;

	HGLOBAL hnew;
	LPSTR	pstr = (LPSTR)GlobalLock(hglobalA);
	DWORD	dwSize = GlobalSize(hglobalA);
	LONG	cbSize = (dwSize + 1) * sizeof(WCHAR);
	
    hnew = GlobalAlloc(GMEM_FIXED, cbSize);
	if(hnew)
	{
		LPWSTR pwstr = (LPWSTR)GlobalLock(hnew);
		UnicodeFromMbcs(pwstr, dwSize + 1, pstr);
		GlobalUnlock(hnew);		
	}
	GlobalUnlock(hglobalA);
	return hnew;
}

/*
 *	CW32System::TextHGlobalWtoA(hglobalW)
 *
 *	@func
 *		converts a unicode text hglobal into a newly allocated
 *		allocated hglobal with ANSI data
 *
 *	@devnote
 *		does *NOT* free the incoming hglobal
 */
HGLOBAL CW32System::TextHGlobalWtoA(
	HGLOBAL hglobalW )
{
	TRACEBEGIN(TRCSUBSYSUTIL, TRCSCOPEINTERN, "CW32System::TextHGlobalWtoA");

	if(!hglobalW)
		return NULL;

	HGLOBAL hnew = NULL;
	LPWSTR 	pwstr = (LPWSTR)GlobalLock(hglobalW);
	DWORD	dwSize = GlobalSize(hglobalW);
	LONG	cbSize = (dwSize * 2) * sizeof(CHAR);
	hnew = GlobalAlloc(GMEM_FIXED, cbSize);

	if( hnew )
	{
		LPSTR pstr = (LPSTR)GlobalLock(hnew);
		MbcsFromUnicode(pstr, cbSize, pwstr );
		GlobalUnlock(hnew);
	}
	GlobalUnlock(hglobalW);
	return hnew;
}	

/*
 *	CW32System::ConvertLanguageIDtoCodePage (lid)
 *
 *	@mfunc		Maps a language ID to a Code Page
 *
 *	@rdesc		returns Code Page
 *
 *	@devnote:
 *		This routine takes advantage of the fact that except for Chinese,
 *		the code page is determined uniquely by the primary language ID,
 *		which is given by the low-order 10 bits of the lcid.
 *
 *		The WORD CodePageTable could be replaced by a BYTE with the addition
 *		of a couple of if's and the BYTE table replaced by a nibble table
 *		with the addition of a shift and a mask.  Since the table is only
 *		96 bytes long, it seems that the simplicity of using actual code page
 *		values is worth the extra bytes.
 */
UINT CW32System::ConvertLanguageIDtoCodePage(
	WORD lid)				//@parm Language ID to map to code page
{
	UINT j = PRIMARYLANGID(lid);			// j = primary language (PLID)

	if(j >= LANG_CROATIAN)					// PLID = 0x1a
	{
		if (lid == lidSerbianCyrillic ||	// Special case for LID = 0xc1a
			lid == lidAzeriCyrillic	  ||
			lid == lidUzbekCyrillic)
		{
			return 1251;					// Use Cyrillic code page
		}

		if(j >= nCodePageTable)				// Most languages above table
		{									//  take 1252
			if (j == 0x57 || j == 0x61 || IN_RANGE(0x4e, j, 0x4f))
				return CP_DEVANAGARI;	

			if (j == 0x49)
				return CP_TAMIL;

			return 1252;
		}
	}

	j = CodePageTable[j];					// Translate PLID to code page

	if(j != 950 || lid == LANG_TAIWAN)		// All but China (except for Taiwan)
		return j;

	return 936;								// Hong Kong SAR, Singapore, and PRC
}

/*
 *	CW32System::GetLocaleLCID ()
 *
 *	@mfunc		Maps an LCID for thread to a Code Page
 *
 *	@rdesc		returns Code Page
 */
LCID CW32System::GetLocaleLCID()
{
	return GetThreadLocale();
}

/*
 *	CW32System::GetLocaleCodePage ()
 *
 *	@mfunc		Maps an LCID for thread to a Code Page
 *
 *	@rdesc		returns Code Page
 */
UINT CW32System::GetLocaleCodePage()
{
#ifdef DEBUG
	UINT cpg = W32->DebugDefaultCpg();
	if (cpg)
		return cpg;
#endif
	return W32->ConvertLanguageIDtoCodePage(GetThreadLocale());
}

/*
 *	CW32System::GetKeyboardLCID ()
 *
 *	@mfunc		Gets LCID for keyboard active on current thread
 *
 *	@rdesc		returns Code Page
 */
LCID CW32System::GetKeyboardLCID(DWORD dwMakeAPICall)
{
	return (WORD)GetKeyboardLayout(dwMakeAPICall);
}

/*
 *	CW32System::GetKeyboardCodePage ()
 *
 *	@mfunc		Gets Code Page for keyboard active on current thread
 *
 *	@rdesc		returns Code Page
 */
UINT CW32System::GetKeyboardCodePage(DWORD dwMakeAPICall)
{
	return W32->ConvertLanguageIDtoCodePage((WORD)GetKeyboardLayout(dwMakeAPICall));
}

/*
 *	CW32System::InitKeyboardFlags ()
 *
 *	@mfunc
 *		Initializes keyboard flags. Used when control gains focus. Note that
 *		Win95 doesn't support VK_RSHIFT, so if either shift key is pressed
 *		when focus is regained, it'll be assumed to be the left shift.
 */
void CW32System::InitKeyboardFlags()
{
	_wKeyboardFlags = 0;
	if(GetKeyState(VK_SHIFT) < 0)
		SetKeyboardFlag(GetKeyState(VK_RSHIFT) < 0 ? RSHIFT : LSHIFT);
}

/*
 *	CW32System::GetKeyboardFlag (dwKeyMask, wKey)
 *
 *	@mfunc
 *		Return whether wKey is depressed. Check with OS for agreement.
 *		If OS says it isn't depressed, reset our internal flags. In
 *		any event, return TRUE/FALSE in agreement with the system (bad
 *		client may have eaten keystrokes, thereby destabilizing our
 *		internal keyboard state.
 *
 *	@rdesc
 *		TRUE iff wKey is depressed
 */
BOOL CW32System::GetKeyboardFlag (
	WORD dwKeyMask,	//@parm _wKeyboardFlags mask like ALT, CTRL, or SHIFT
	WORD wKey)		//@parm VK_xxx like VK_MENU, VK_CONTROL, or VK_SHIFT
{
	BOOL fFlag = (GetKeyboardFlags() & dwKeyMask) != 0;

	if(fFlag ^ ((GetKeyState(wKey) & 0x8000) != 0))
	{	
		// System doesn't agree with our internal state
		// (bad client ate a WM_KEYDOWN)
		if(fFlag)
		{
			ResetKeyboardFlag(dwKeyMask);
			return FALSE;					
		}
		// Don't set an internal _wKeyboardFlag since we check for it
		// anyhow and client might not send WM_KEYUP either
		return TRUE;
	}							
	return fFlag;
}

/*
 *	CW32System::IsAlef(ch)
 *
 *	@func
 *		Used to determine if base character is a Arabic-type Alef.
 *
 *	@rdesc
 *		TRUE iff the base character is an Arabic-type Alef.
 *
 *	@comm
 *		AlefWithMaddaAbove, AlefWithHamzaAbove, AlefWithHamzaBelow,
 *		and Alef are valid matches.
 */
BOOL CW32System::IsAlef(
	TCHAR ch)
{
	return IN_RANGE(0x622, ch, 0x627) && ch != 0x624 && ch != 0x626;
}

/*
 *	CW32System::IsBiDiLcid(lcid)
 *
 *	@func
 *		Return TRUE if lcid corresponds to an RTL language
 *
 *	@rdesc
 *		TRUE if lcid corresponds to an RTL language
 */
BOOL CW32System::IsBiDiLcid(
	LCID lcid)
{
	return
		PRIMARYLANGID(lcid) == LANG_ARABIC ||
		PRIMARYLANGID(lcid) == LANG_HEBREW ||
		PRIMARYLANGID(lcid) == LANG_URDU ||
		PRIMARYLANGID(lcid) == LANG_FARSI;
}

/*
 *	CW32System::IsIndicLcid(lcid)
 *
 *	@func
 *		Return TRUE if lcid corresponds to an Indic language
 *
 *	@rdesc
 *		TRUE if lcid corresponds to an Indic language
 */
BOOL CW32System::IsIndicLcid(
	LCID lcid)
{
	WORD	wLangId = PRIMARYLANGID(lcid);
	
	return
		wLangId == LANG_HINDI 	||
		wLangId == LANG_KONKANI ||
		wLangId == LANG_NEPALI 	||
		IN_RANGE(LANG_BENGALI, wLangId, LANG_SANSKRIT);
}

/*
 *	CW32System::IsIndicKbdInstalled()
 *
 *	@func
 *		Return TRUE if any Indic kbd installed
 */
bool CW32System::IsIndicKbdInstalled()
{
	for (int i = INDIC_FIRSTINDEX; i <= INDIC_LASTINDEX; i++)
		if (_hkl[i] != 0)
			return true;
	return false;
}

/*
 *	CW32System::IsComplexScriptLcid(lcid)
 *
 *	@func
 *		Return TRUE if lcid corresponds to any complex script locales
 *
 */
BOOL CW32System::IsComplexScriptLcid(
	LCID lcid)
{
	return	IsBiDiLcid(lcid) ||
			PRIMARYLANGID(LANGIDFROMLCID(lcid)) == LANG_THAI ||
			PRIMARYLANGID(LANGIDFROMLCID(lcid)) == LANG_VIETNAMESE ||
			IsIndicLcid(lcid);
}

/*
 *	CW32System::IsBiDiDiacritic(ch)
 *
 *	@func	Used to determine if character is a Arabic or Hebrew diacritic.
 *
 *  @rdesc  TRUE iff character is a diacritic
 */
BOOL CW32System::IsBiDiDiacritic(
	TCHAR ch)
{
	return IN_RANGE(0x64B, ch, 0x670) && (ch <= 0x652 || ch == 0x670) ||	// Arabic
		   IN_RANGE(0x591, ch, 0x5C4) && (ch != 0x5A2 && ch != 0x5BA &&		// Hebrew
				ch != 0x5BE && ch != 0x5C0 && ch != 0x5C3);					
}


/*
 *	CW32System::IsVietCdmSequenceValid(ch1, ch2)
 *
 *	@mfunc
 *		Check if ch2 can follow ch1 in case ch2 is a combining diacritic mark (CDM).
 *		Main use is for Vietnamese users (Chau Vu provides the logic below).
 *
 *	@rdesc
 *		TRUE if ch2 can follow ch1
 */
BOOL CW32System::IsVietCdmSequenceValid(
	WCHAR ch1,
	WCHAR ch2)
{
	if (!IN_RANGE(0x300, ch2, 0x323) ||		// Fast out
		!IN_RANGE(0x300, ch2, 0x301) && ch2 != 0x303 && ch2 != 0x309 && ch2 != 0x323)
	{
		return TRUE;						// Not Vietnamese tone mark
	}
	//								ô	  ê		â
	static const BYTE vowels[] = {0xF4, 0xEA, 0xE2, 'y', 'u', 'o', 'i', 'e', 'a'};

	for(int i = ARRAY_SIZE(vowels); i--; )
		if((ch1 | 0x20) == vowels[i])		// Vietnamese tone mark follows
			return TRUE;					//  vowel

	return IN_RANGE(0x102, ch1, 0x103) ||	// A-breve, a-breve
		   IN_RANGE(0x1A0, ch1, 0x1A1) ||	// O-horn,  o-horn
		   IN_RANGE(0x1AF, ch1, 0x1B0);		// U-horn,  u-horn
}

/*
 *  CW32System::IsFELCID(lcid)
 *
 *	@mfunc
 *		Returns TRUE iff lcid is for a East Asian country/region.
 *
 *	@rdesc
 *		TRUE iff lcid is for a East Asian country/region.
 */
bool CW32System::IsFELCID(
	LCID lcid)
{
	switch(PRIMARYLANGID(LANGIDFROMLCID(lcid)))
	{
		case LANG_CHINESE:
		case LANG_JAPANESE:
		case LANG_KOREAN:
			return true;
	}
	return false;
}

/*
 *  CW32System::IsFECharSet(bCharSet)
 *
 *	@mfunc
 *		Returns TRUE iff charset may be for a East Asian country/region.
 *
 *	@rdesc
 *		TRUE iff charset may be for a East Asian country/region.
 *
 */
BOOL CW32System::IsFECharSet(
	BYTE bCharSet)
{
	switch(bCharSet)
	{
		case CHINESEBIG5_CHARSET:
		case SHIFTJIS_CHARSET:
		case HANGEUL_CHARSET:
		case JOHAB_CHARSET:
		case GB2312_CHARSET:
			return TRUE;
	}

	return FALSE;
}

/*
 *  CW32System::Is8BitCodePage(CodePage)
 *
 *	@mfunc
 *		Returns TRUE iff the codepage is 8-bit
 */
BOOL CW32System::Is8BitCodePage(
	unsigned CodePage)
{
	if(!CodePage)
		CodePage = GetACP();

	return IN_RANGE(1250, CodePage, 1258) || CodePage == 874;
}

/*
 *  CW32System::IsFECodePageFont(dwFontCodePageSig)
 *
 *	@mfunc
 *		Returns TRUE iff the font codepage signature reveals only East Asian support
 */
BOOL CW32System::IsFECodePageFont(
	DWORD dwFontCodePageSig)
{
	DWORD	dwFE 	= 0x001e0000;	// Shift-JIS + PRC + Hangeul + Taiwan
	DWORD	dwOthers = 0x000101fc;	// The rest of the world except for Latin-1 and Latin-2

	return (dwFontCodePageSig & dwFE) && !(dwFontCodePageSig & dwOthers);
}

/*
 *  CW32System::IsRTLChar(ch)
 *
 *	@mfunc
 *		Returns TRUE iff ch Arabic or Hebrew
 *
 *	@rdesc
 *		TRUE iff ch is Arabic or Hebrew
 */
BOOL IsRTLChar(
	TCHAR	ch)
{
	// Remark: what about Arabic Presentation Forms?
	//  (0xFB50 - 0xFDFF, 0xFE70 - 0xFEFF)

	return IN_RANGE(0x590, ch, 0x6FF) || ch == RTLMARK;
}

/*
 *  CW32System::IsRTLCharSet(bCharSet)
 *
 *	@mfunc
 *		Returns TRUE iff charset is Arabic or Hebrew
 *
 *	@rdesc
 *		TRUE iff charset may be for Arabic or Hebrew
 */
BOOL CW32System::IsRTLCharSet(
	BYTE bCharSet)
{
	return IN_RANGE(HEBREW_CHARSET, bCharSet, ARABIC_CHARSET);
}

/*
 *	CW32System::IsZWG(ch, bCharSet)
 *
 *	@mfunc
 *		Determines whether a single-byte character is a zero-width glyph
 *		or not. Requires a charset.
 *
 *	@rdesc
 *		TRUE if the character is a zero-width glyph, FALSE otherwise.
 */
BOOL CW32System::IsZWG(
	char ch,
	BYTE bCharSet)
{
	if(IN_RANGE(0x9D, ch, 0x9E))
		return bCharSet == ARABIC_CHARSET;

	if(IN_RANGE(0xFD, ch, 0xFE))
		return IN_RANGE(HEBREW_CHARSET, bCharSet, ARABIC_CHARSET);

	return FALSE;
}

typedef struct {
	WCHAR codepoint;
	WORD  CharFlags;
	BYTE  runlength;
} Data_125X;

/*
 *  CW32System::GetCharFlags125x(ch)
 *
 *	@mfunc
 *		Returns char flags for ch as defined in rgCpgCharSet for 125x
 *		codepages. Bit 0: 1252, bit 1: 1250, bit 2: 1251, else bit x:
 *		125x (for 1253 - 1258).
 *
 *	@rdesc
 *		125x char flags for ch
 */
DWORD CW32System::GetCharFlags125x(
	WCHAR	ch)			//@parm Char to examine
{
	static const WORD rgCpgMask[] = {
		0x1FF,		// 0xA0
		0x131,		// 0xA1
		0x1F1,		// 0xA2
		0x1F9,		// 0xA3
		0x1DF,		// 0xA4
		0x179,		// 0xA5
		0x1FF,		// 0xA6
		0x1FF,		// 0xA7
		0x1FB,		// 0xA8
		0x1FF,		// 0xA9
		0x111,		// 0xAA
		0x1FF,		// 0xAB
		0x1FF,		// 0xAC
		0x1FF,		// 0xAD
		0x1FF,		// 0xAE
		0x1F1,		// 0xAF
		0x1FF,		// 0xB0
		0x1FF,		// 0xB1
		0x1F9,		// 0xB2
		0x1F9,		// 0xB3
		0x1F3,		// 0xB4
		0x1FF,		// 0xB5
		0x1FF,		// 0xB6
		0x1FF,		// 0xB7
		0x1F3,		// 0xB8
		0x1F1,		// 0xB9
		0x111,		// 0xBA
		0x1FF,		// 0xBB
		0x1F1,		// 0xBC
		0x1F9,		// 0xBD
		0x1F1,		// 0xBE
		0x131,		// 0xBF
		0x111,		// 0xC0
		0x113,		// 0xC1
		0x113,		// 0xC2
		0x011,		// 0xC3
		0x193,		// 0xC4
		0x191,		// 0xC5
		0x191,		// 0xC6
		0x113,		// 0xC7
		0x111,		// 0xC8
		0x193,		// 0xC9
		0x111,		// 0xCA
		0x113,		// 0xCB
		0x011,		// 0xCC
		0x113,		// 0xCD
		0x113,		// 0xCE
		0x111,		// 0xCF
		0x001,		// 0xD0
		0x111,		// 0xD1
		0x011,		// 0xD2
		0x193,		// 0xD3
		0x113,		// 0xD4
		0x091,		// 0xD5
		0x193,		// 0xD6
		0x1F3,		// 0xD7
		0x191,		// 0xD8
		0x111,		// 0xD9
		0x113,		// 0xDA
		0x111,		// 0xDB
		0x193,		// 0xDC
		0x003,		// 0xDD
		0x001,		// 0xDE
		0x193,		// 0xDF
		0x151,		// 0xE0
		0x113,		// 0xE1
		0x153,		// 0xE2
		0x011,		// 0xE3
		0x193,		// 0xE4
		0x191,		// 0xE5
		0x191,		// 0xE6
		0x153,		// 0xE7
		0x151,		// 0xE8
		0x1D3,		// 0xE9
		0x151,		// 0xEA
		0x153,		// 0xEB
		0x011,		// 0xEC
		0x113,		// 0xED
		0x153,		// 0xEE
		0x151,		// 0xEF
		0x001,		// 0xF0
		0x111,		// 0xF1
		0x011,		// 0xF2
		0x193,		// 0xF3
		0x153,		// 0xF4
		0x091,		// 0xF5
		0x193,		// 0xF6
		0x1F3,		// 0xF7
		0x191,		// 0xF8
		0x151,		// 0xF9
		0x113,		// 0xFA
		0x151,		// 0xFB
		0x1D3,		// 0xFC
		0x003,		// 0xFD
		0x001,		// 0xFE
		0x111		// 0xFF
	};
	static const Data_125X Table_125X[] = {
		{ 0x100, 0x080,  2},
		{ 0x102, 0x102,  2},
		{ 0x104, 0x082,  4},
		{ 0x10c, 0x082,  2},
		{ 0x10e, 0x002,  2},
		{ 0x110, 0x102,  2},
		{ 0x112, 0x080,  2},
		{ 0x116, 0x080,  2},
		{ 0x118, 0x082,  2},
		{ 0x11a, 0x002,  2},
		{ 0x11e, 0x010,  2},
		{ 0x122, 0x080,  2},
		{ 0x12a, 0x080,  2},
		{ 0x12e, 0x080,  2},
		{ 0x130, 0x010,  2},
		{ 0x136, 0x080,  2},
		{ 0x139, 0x002,  2},
		{ 0x13b, 0x080,  2},
		{ 0x13d, 0x002,  2},
		{ 0x141, 0x082,  4},
		{ 0x145, 0x080,  2},
		{ 0x147, 0x002,  2},
		{ 0x14c, 0x080,  2},
		{ 0x150, 0x002,  2},
		{ 0x152, 0x151,  2},
		{ 0x154, 0x002,  2},
		{ 0x156, 0x080,  2},
		{ 0x158, 0x002,  2},
		{ 0x15a, 0x082,  2},
		{ 0x15e, 0x012,  2},
		{ 0x160, 0x093,  2},
		{ 0x162, 0x002,  4},
		{ 0x16a, 0x080,  2},
		{ 0x16e, 0x002,  4},
		{ 0x172, 0x080,  2},
		{ 0x178, 0x111,  1},
		{ 0x179, 0x082,  4},
		{ 0x17d, 0x083,  2},
		{ 0x192, 0x179,  1},
		{ 0x1A0, 0x100,  2},
		{ 0x1AF, 0x100,  2},
		{ 0x2c6, 0x171,  1},
		{ 0x2c7, 0x082,  1},
		{ 0x2d8, 0x002,  1},
		{ 0x2d9, 0x082,  1},
		{ 0x2db, 0x082,  1},
		{ 0x2dc, 0x131,  1},
		{ 0x2dd, 0x002,  1},
		{ 0x300, 0x100,  2},
		{ 0x303, 0x100,  1},
		{ 0x309, 0x100,  1},
		{ 0x323, 0x100,  1},
		{ 0x384, 0x008,  3},
		{ 0x388, 0x008,  3},
		{ 0x38c, 0x008,  1},
		{ 0x38e, 0x008, 20},
		{ 0x3a3, 0x008, 44},
		{ 0x401, 0x004, 12},
		{ 0x40e, 0x004, 66},
		{ 0x451, 0x004, 12},
		{ 0x45e, 0x004,  2},
		{ 0x490, 0x004,  2},
		{ 0x5b0, 0x020, 20},
		{ 0x5d0, 0x020, 27},
		{ 0x5F0, 0x020,  5},
		{ 0x60c, 0x040,  1},
		{ 0x61b, 0x040,  1},
		{ 0x61f, 0x040,  1},
		{ 0x621, 0x040, 26},
		{ 0x640, 0x040, 19},
		{ 0x679, 0x040,  1},
		{ 0x67e, 0x040,  1},
		{ 0x686, 0x040,  1},
		{ 0x688, 0x040,  1},
		{ 0x691, 0x040,  1},
		{ 0x698, 0x040,  1},
		{ 0x6a9, 0x040,  1},
		{ 0x6af, 0x040,  1},
		{ 0x6ba, 0x040,  1},
		{ 0x6be, 0x040,  1},
		{ 0x6c1, 0x040,  1},
		{ 0x6d2, 0x040,  1},
		{0x200c, 0x040,  2},
		{0x200e, 0x060,  2},
		{0x2013, 0x1ff,  2},
		{0x2015, 0x008,  1},
		{0x2018, 0x1ff,  3},
		{0x201c, 0x1ff,  3},
		{0x2020, 0x1ff,  3},
		{0x2026, 0x1ff,  1},
		{0x2030, 0x1ff,  1},
		{0x2039, 0x1ff,  2},
		{0x20AA, 0x020,  1},
		{0x20AB, 0x100,  1},
		{0x20AC, 0x1ff,  1},
		{0x2116, 0x004,  1},
		{0x2122, 0x1ff,  1}
	};

	// Easy check for ASCII
	if(ch <= 0x7f)
		return 0x1ff00;

	// Easy check for missing codes
	if(ch > 0x2122)
		return 0;

	if(IN_RANGE(0xA0, ch, 0xFF))
		return rgCpgMask[ch - 0xA0] << 8;

	// Perform binary search to find entry in table
	int low = 0;
	int high = ARRAY_SIZE(Table_125X) - 1;
	int middle;
	int midval;
	int runlength;
	
	while(low <= high)
	{
		middle = (high + low) / 2;
		midval = Table_125X[middle].codepoint;
		if(midval > ch)
			high = middle - 1;
		else
			low = middle + 1;
	
		runlength = Table_125X[middle].runlength;
		if(ch >= midval && ch <= midval + runlength - 1)
			return Table_125X[middle].CharFlags << 8;
	}
	return 0;
}

/*
 *	CW32System::IsUTF8BOM(pstr)
 *
 *	@mfunc
 *		Return TRUE if pstr points at a UTF-8 BOM
 *
 *	@rdesc
 *		TRUE iff pstr points at a UTF-8 BOM
 */
BOOL CW32System::IsUTF8BOM(
	BYTE *pstr)
{
	BYTE *pstrUtf8BOM = szUTF8BOM;

	for(LONG i = 3; i--; )
		if(*pstr++ != *pstrUtf8BOM++)
			return FALSE;
	return TRUE;
}

/*
 *	CW32System::GetTrailBytesCount(ach, cpg)
 *
 *	@mfunc
 *		Returns number of trail bytes iff the byte ach is a lead byte for the code page cpg.
 *
 *	@rdesc
 *		count of trail bytes if ach is lead byte for cpg
 *
 *	@comm
 *		This is needed to support CP_UTF8 as well as DBC.
 *		This function potentially doesn't support as many code pages as the
 *		Win32 IsDBCSLeadByte() function (and it might not be as up-to-date).
 *		An AssertSz() is included to compare the results when the code page
 *		is supported by the system.
 *
 *		Reference: \\sparrow\sysnls\cptable\win95. See code-page txt files
 *		in subdirectories windows\txt and others\txt.
 */
int CW32System::GetTrailBytesCount(BYTE ach, UINT cpg)
{

	if(ach < 0x81)									// Smallest known lead
		return 0;									//  byte = 0x81:
													//  early out
	
	BOOL	bDBLeadByte = FALSE;					// Variable to check
													//  result with system
													//  ifdef DEBUG

	if (cpg == CP_UTF8)
	{
		int	cTrailBytes = 0;						// Number of trail bytes for CP_UTF8(0 - 3)
		
		if (ach >= 0x0F0)							// Handle 4-byte form for 16 UTF-16 planes
			cTrailBytes = 3;						// above the BMP) expect:
													// 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb
		else if (ach >= 0x0E0)						// Need at least 3 bytes of form
			cTrailBytes = 2;						// 1110bbbb 10bbbbbb 10bbbbbb
		else if (ach >= 0x0C0)						// Need at least 2 bytes of form
			cTrailBytes = 1;						// 110bbbbb 10bbbbbb

		return cTrailBytes;
	}
	else if(cpg > 950)								
	{
		if(cpg < 1361)								// E.g., the 125x's are
			return 0;								//  SBCSs: early out

		else if(cpg == 1361)								// Korean Johab
			bDBLeadByte = IN_RANGE(0x84, ach, 0xd3) ||		// 0x84 <= ach <= 0xd3
				   IN_RANGE(0xd8, ach, 0xde) ||				// 0xd8 <= ach <= 0xde
				   IN_RANGE(0xe0, ach, 0xf9);				// 0xe0 <= ach <= 0xf9

		else if(cpg == 10001)						// Mac Japanese
			goto JIS;

		else if(cpg == 10002)						// Mac Trad Chinese (Big5)
			bDBLeadByte = ach <= 0xfe;

		else if(cpg == 10003)						// Mac Korean
			bDBLeadByte = IN_RANGE(0xa1, ach, 0xac) ||		// 0xa1 <= ach <= 0xac
				   IN_RANGE(0xb0, ach, 0xc8) ||		// 0xb0 <= ach <= 0xc8
				   IN_RANGE(0xca, ach, 0xfd);		// 0xca <= ach <= 0xfd

		else if(cpg == 10008)						// Mac Simplified Chinese
			bDBLeadByte = IN_RANGE(0xa1, ach, 0xa9) ||		// 0xa1 <= ach <= 0xa9
				   IN_RANGE(0xb0, ach, 0xf7);		// 0xb0 <= ach <= 0xf7
	}
	else if (cpg >= 932)							// cpg <= 950
	{
		if(cpg == 950 || cpg == 949 || cpg == 936)	// Chinese (Taiwan, HK),
			bDBLeadByte = ach <= 0xfe;						//  Korean Ext Wansung,
													//  PRC GBK: 0x81 - 0xfe
		else if(cpg == 932)							// Japanese
JIS:		bDBLeadByte = ach <= 0x9f || IN_RANGE(0xe0, ach, 0xfc);
	}

	#ifdef DEBUG
	TCHAR	ch;
	static	BYTE asz[2] = {0xe0, 0xe0};				// All code pages above

	// if cpg == 0, fRet will FALSE but IsDBCSLeadByteEx may succeed.
	if ( cpg && cpg != CP_SYMBOL && cpg != CP_UTF8)
	{
		// If system supports cpg, then fRet should agree with system result
		AssertSz(MultiByteToWideChar(cpg, 0, (char *)asz, 2, &ch, 1) <= 0 ||
			bDBLeadByte == IsDBCSLeadByteEx(cpg, ach),
			"bDBLeadByte differs from IsDBCSLeadByteEx()");
	}
	#endif

	return bDBLeadByte ? 1 : 0;
}

/*
 *	CW32System::GetGdiCharSet(bCharSet)
 *
 *	@func
 *		Map CHARFORMAT charset to GDI charset (charset used in CreateFont)
 */
BYTE CW32System::GetGdiCharSet(BYTE	bCharSet)
{
	return IsPrivateCharSet(bCharSet) ? DEFAULT_CHARSET : bCharSet;
}

/*
 *	CW32System::GetCharSet(nCP, pScriptIndex)
 *
 *	@func
 *		Get character set for code page <p nCP>. Also returns script index
 *		in *pScriptIndex
 *
 *	@rdesc
 *		CharSet for code page <p nCP>
 */
BYTE CW32System::GetCharSet(
	INT  nCP,				//@parm Code page or index
	int *pScriptIndex)		//@parm Out parm to receive index
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "GetCharSet");

	if(nCP < NCHARSETS)					// nCP is already an index
	{
		nCP = max(nCP, 0);
		if(pScriptIndex)
			*pScriptIndex = nCP;
		return rgCpgCharSet[nCP].bCharSet;
	}

	const CPGCHAR *	pcpgchar = rgCpgCharSet;

	for (int i = 0; i < cCpgCharSet; i++, ++pcpgchar)
	{
		if(pcpgchar->nCodePage == nCP)
			break;
	}
	if(i == cCpgCharSet)
		i = -1;

	if (pScriptIndex)
		*pScriptIndex = i;

	return i >= 0 ? pcpgchar->bCharSet : 0;
}

/*
 *	CW32System::MatchFECharSet(dwCharInfo, dwFontSig)
 *
 *	@func
 *		Get a FE character set for a FE char
 *
 *	@rdesc
 *		CharSet
 */
BYTE CW32System::MatchFECharSet(
	DWORD  dwCharInfo,		//@parm Char Info
	DWORD  dwFontSig)		//@parm Font Signature
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "CW32System::MatchFECharSet");

	// Convert Font Signature to our internal char info
	dwFontSig = (dwFontSig << 8) & ~0x00ffff;

	// Get rid of lower word
	dwCharInfo &= ~0x00ffff;

	if (dwCharInfo & dwFontSig & (fKANA | fCHINESE | fHANGUL | fBIG5)) // Perfect match
		goto Exit;											

	if (!(dwFontSig & (fKANA | fCHINESE | fHANGUL | fBIG5)))	// Not a FE font
		goto Exit;

	if (dwCharInfo & (fCHINESE | fBIG5))
	{
		if (dwFontSig & fBIG5)
			return CHINESEBIG5_CHARSET;
		if (dwFontSig & fHANGUL)
			return HANGEUL_CHARSET;
		if (dwFontSig & fKANA)
			return SHIFTJIS_CHARSET;
	}

Exit:
	return GetCharSet(W32->ScriptIndexFromFontSig(dwCharInfo >> 8), NULL);
}

/*
 *	CW32System::GetCodePage(bCharSet)
 *
 *	@func
 *		Get code page for character set <p bCharSet>
 *
 *	@rdesc
 *		Code page for character set <p bCharSet>
 */
INT CW32System::GetCodePage(
	BYTE bCharSet)		//@parm CharSet
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "GetCodePage");

	const CPGCHAR *pcpgchar = rgCpgCharSet;

	for (int i = 0; i < cCpgCharSet ; i++, ++pcpgchar)
	{
		if (pcpgchar->bCharSet == bCharSet)
			return pcpgchar->nCodePage;
	}
	return 0;
}

/*
 *	CW32System::GetFontSig(wCharSet)
 *
 *	@func
 *		Get font signature bits for character set <p wCharSet>. If
 *		wCharSet > 256, treat as a codepage
 *
 *	@rdesc
 *		Font signature mask for character set (or codepage) <p bCharSet>
 */
DWORD CW32System::GetFontSig(
	WORD wCharSet)
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "CW32System::GetFontSig");

	const CPGCHAR *	pcpgchar = rgCpgCharSet;
	BOOL fCodePage = wCharSet > 255;

	for (int i = 0; i < cCpgCharSet; i++)
	{
		if(wCharSet == (WORD)(fCodePage ? pcpgchar->nCodePage : pcpgchar->bCharSet))
			return pcpgchar->dwFontSig;
		++pcpgchar;
	}
	return 0;
}

/*
 *	CW32System::GetFirstAvailCharSet(DWORD dwFontSig)
 *
 *	@func
 *		Get first available charset from font signature bits
 */
BYTE CW32System::GetFirstAvailCharSet(
	DWORD dwFontSig)
{
	const CPGCHAR *	pcpgchar = rgCpgCharSet;

	for (int i = 0; i < cCpgCharSet; i++)
	{
		if(pcpgchar->dwFontSig & dwFontSig)
			return pcpgchar->bCharSet;
		++pcpgchar;
	}
	return ANSI_CHARSET;
}

/*
 *	CW32System::GetFontSigFromScript(iScript)
 *
 *	@func
 *		Get font signature bits for script <p iScript>
 *
 *	@rdesc
 *		Font signature mask for script <p iScript>
 */
DWORD CW32System::GetFontSigFromScript(
	int iScript)
{
	return rgCpgCharSet[iScript].dwFontSig;
}

/*
 *	CW32System::ScriptIndexFromFontSig(dwFontSig)
 *
 *	@func
 *		Get script index from font signature
 *
 *	@rdesc
 *		Script index from font signature <p dwFontSig>
 *
 *	@devnote
 *		Linear search
 */
LONG CW32System::ScriptIndexFromFontSig(
	DWORD dwFontSig)
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "CW32System::ScriptIndexFromFontSig");

	const CPGCHAR *pcpgchar = rgCpgCharSet;

	for (int i = 0; i < cCpgCharSet; i++, ++pcpgchar)
	{
		if(pcpgchar->dwFontSig == dwFontSig)
			return i;
	}
	return -1;								// Not found
}

/*
 *	CW32System::ScriptIndexFromCharSet(bCharSet)
 *
 *	@func
 *		Get script index from bCharSet
 *
 *	@rdesc
 *		Script index character set <p bCharSet>
 *
 *	@devnote
 *		Linear search
 */
LONG CW32System::ScriptIndexFromCharSet(
	BYTE bCharSet)
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "CW32System::ScriptIndexFromCharSet");

	const CPGCHAR *pcpgchar = rgCpgCharSet;

	for (int i = 0; i < cCpgCharSet; i++, ++pcpgchar)
	{
		if(pcpgchar->bCharSet == bCharSet)
			return i;
	}
	return -1;								// Not found
}

/*
 *	CW32System::IsCharSetValid(bCharSet)
 *
 *	@func
 *		Return TRUE iff <p bCharSet> is a valid character set index
 *
 *	@rdesc
 *		TRUE iff <p bCharSet> is a valid character set index
 */
BOOL CW32System::IsCharSetValid(
	BYTE bCharSet)		//@parm CharSet
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "CW32System::IsCharSetValid");

	return ScriptIndexFromCharSet(bCharSet) >= 0;
}

/*
 *	CW32System::GetScreenDC(void)
 *
 *	@mfunc
 *	Returns you a default screen DC which richedit caches for its lifetime.
 *
 *	Note, you need to serialize access to DCs, so make sure its used in the
 *	renderer and measurer or otherwise protected by a CLock.
 *
 *	@rdesc
 *		Screen HDC if succeeded.
 */
HDC	CW32System::GetScreenDC()
{
	if (!_hdcScreen)
		_hdcScreen = CreateIC(L"DISPLAY", NULL, NULL, NULL);;

	//Verify DC validity
	Assert(GetDeviceCaps(_hdcScreen, LOGPIXELSX));

	return _hdcScreen;
}

/*
 *	CW32System::GetTextMetrics(hdc, &lf, &tm)
 *
 *	@mfunc
 *		CreateFontIndirect(lf), select into hdc, and get TEXTMETRICS
 *
 *	@rdesc
 *		TRUE if succeeded and selected facename is same as lf.lfFaceName
 */
BOOL CW32System::GetTextMetrics(
	HDC			hdc,
	LOGFONT &	lf,
	TEXTMETRIC &tm)
{
	HFONT hfont = CreateFontIndirect(&lf);
	if(!hfont)
		return FALSE;

	HFONT hfontOld = SelectFont(hdc, hfont);
	WCHAR szFaceName[LF_FACESIZE + 1];

	BOOL  fRet = GetTextFace(hdc, LF_FACESIZE, szFaceName) &&
				 !wcsicmp(lf.lfFaceName, szFaceName) &&
				  W32->GetTextMetrics(hdc, &tm);

	SelectFont(hdc, hfontOld);
	DeleteObject(hfont);
	return fRet;
}

/*
 *	CW32System::ValidateStreamWparam(wparam)
 *
 *	@mfunc
 *		Examine lparam to see if hiword is a valid codepage. If not set it
 *		to 0 and turn off SF_USECODEPAGE flag
 *
 *	@rdesc
 *		Validated lparam
 */
WPARAM CW32System::ValidateStreamWparam(
	WPARAM wparam)		//@parm EM_STREAMIN/OUT wparam
{
	TRACEBEGIN(TRCSUBSYSRTFR, TRCSCOPEINTERN, "CW32System::ValidateStreamWparam");

	if ((wparam & SF_USECODEPAGE) && !IsValidCodePage(HIWORD(wparam)) &&
		HIWORD(wparam) != CP_UTF8)
	{
		// Invalid codepage, so reset codepage parameters
		wparam &= 0xFFFF & ~SF_USECODEPAGE;
	}
	return wparam;
}

/*
 *	CW32System::MECharClass(ch)
 *
 *	@func
 *		return ME character type for purposes of CharSet stamping.  Values
 *		are:
 *
 *		0: Arabic (specific RTL)
 *		1: Hebrew (specific RTL)
 *		2: RTL (generic RTL, e.g., RTL mark)
 *		3: LTR
 *		4: EOP or start/end of text
 *		5: ASCII digit
 *		6: punctuation and neutrals
 *
 *	@rdesc
 *		ME character class
 */
CC CW32System::MECharClass(
	TCHAR	ch)
{
	AssertSz(CC_NEUTRAL > CC_ASCIIDIGIT && CC_ASCIIDIGIT > CC_EOP &&
			 CC_EOP > CC_LTR && CC_LTR > CC_RTL && CC_RTL > CC_ARABIC,
		"CW32System::MECharClass: invalid CC values");

	// Work down Unicode values from large to small. Use nested if
	// statements to reduce the number executed.

	// Remark: what about Arabic Presentation Forms?
	//  (0xFB50 - 0xFDFF, 0xFE70 - 0xFEFF)

	if(ch >= 0x700)
	{
		if(IN_RANGE(ENQUAD, ch, RTLMARK))
		{								// ENQUAD thru RTLMARK
			if(ch == RTLMARK)			// Maybe add more Unicode general
				return CC_RTL;			//  punctuation?

			if(IN_RANGE(ZWNJ, ch, ZWJ))	// ZWNJ & ZWJ are handled as Arabic,
				return CC_ARABIC;		//  even though they actually shouldn't
										//  affect layout.
			if(ch < ZWNJ)
				return CC_NEUTRAL;		// Various blanks are neutral
		}
		return CC_LTR;
	}

	if(ch >= 0x40)
	{
		if(ch >= 0x590)
			return (ch >= 0x600) ? CC_ARABIC : CC_HEBREW;

		if(IN_RANGE(0x7B, (ch | 0x20), 0x7F) || ch == 0x60 || ch == 0x40)
			return CC_NEUTRAL;			// [\]^_{|}~`@

		return CC_LTR;
	}

	if(ch >= 0x20)
	{
		if(IN_RANGE(0x30, ch, 0x39))
			return CC_ASCIIDIGIT;

		return CC_NEUTRAL;
	}

	Assert(ch < 0x20);
	if((1 << ch) & 0x00003201) /* IsASCIIEOP(ch) || ch == TAB || !ch */
		return CC_EOP;
		
	return CC_LTR;	
}

/*
 *	CW32System::ScriptIndexFromChar (ch)
 *
 *	@mfunc
 *		Returns index into CharSet/CodePage table rgCpgCharSet corresponding
 *		to the Unicode character ch provided such an assignment is
 *		reasonably unambiguous, that is, the currently assigned Unicode
 *		characters in various ranges have Windows code-page equivalents.
 *		Ambiguous or impossible assignments return UNKNOWN_INDEX, which
 *		means that the character can only be represented by Unicode in this
 *		simple model.  Note that the UNKNOWN_INDEX, HAN_INDEX, and FE_INDEX
 *		are negative values, i.e., they imply further processing to figure
 *		out what (if any) charset index to use.  Other indices may also
 *		require run processing, such as the blank in BiDi text.  We need
 *		to mark our right-to-left runs with an Arabic or Hebrew char set,
 *		while we mark left-to-right runs with a left-to-right char set.
 *
 *		If the index has the FE_FLAG bit set, then the character is some kind
 *		of FE "DBCS" character, like a Han, Kana, or Hangul character.
 *		FE_INDEX, HAN_INDEX, HANGUL_INDEX, and SHIFTJIS_INDEX all have this
 *		flag set.
 *
 *	@rdesc
 *		Script index in low word, and possible some extra flags, like
 *		FE_FLAG in high word
 */
LONG CW32System::ScriptIndexFromChar(
	TCHAR ch)		//@parm Unicode character to examine
{
	if(ch < 256)
		return ANSI_INDEX;

	if(ch < 0x700)
	{
		if(ch >= 0x600)
			return ARABIC_INDEX;

		if(ch > 0x590)
			return HEBREW_INDEX;

		if(ch < 0x500)
		{
			if(ch >= 0x400)
				return RUSSIAN_INDEX;

			if(ch >= 0x370)
				return GREEK_INDEX;
		}
	}
	else if(ch < 0x2500)
	{	
		if(IN_RANGE(0xE00, ch, 0xE7F))	// Thai
			return THAI_INDEX;
	}
	else if(ch < 0xAC00)
	{
		if(ch >= 0x3400)				// CJK Ideographs
			return HAN_INDEX;

		if(ch >= 0x3100)
			return FE_INDEX;

		if(ch > 0x3040)					// Katakana and Hiragana
			return SHIFTJIS_INDEX;

		if(ch >= 0x3000)
			return FE_INDEX;
	}
	else if(ch < 0xD800)
		return HANGUL_INDEX;

	else if(ch > 0xFF00)
	{
		if(ch < 0xFF65)					// Fullwidth ASCII and halfwidth
			return HAN_INDEX;			//  CJK punctuation

		if(ch < 0xFFA0)					// Halfwidth Katakana
			return SHIFTJIS_INDEX;

		if(ch < 0xFFE0)					// Halfwidth Jamo
			return HANGUL_INDEX;

		if(ch < 0xFFEF)					// Fullwidth punctuation and currency
			return HAN_INDEX;			//  signs; halfwidth forms, arrows
	}									//  and shapes
	return UNKNOWN_INDEX;
}

/*
 *	CW32System::MBTWC (CodePage, dwFlags, pstrMB, cchMB, pstrWC, cchWC, pfNoCodePage)
 *
 *	@mfunc
 *		Convert MultiByte (MB) string pstrMB of length cchMB to WideChar (WC)
 *		string pstrWC of length cchWC according to the flags dwFlags and code
 *		page CodePage.  If CodePage = SYMBOL_CODEPAGE
 *		(usually for SYMBOL_CHARSET strings),
 *		convert each byte in pstrMB to a wide char with a zero high byte
 *		and a low byte equal to the MultiByte string byte, i.e., no
 *		translation other than a zero extend into the high byte.  Else call
 *		the Win32 MultiByteToWideChar() function.
 *
 *	@rdesc
 *		Count of characters converted
 */
int CW32System::MBTWC(
	INT		CodePage,	//@parm Code page to use for conversion
	DWORD	dwFlags,	//@parm Flags to guide conversion
	LPCSTR	pstrMB,		//@parm MultiByte string to convert to WideChar
	int		cchMB,		//@parm Count of chars (bytes) in pstrMB or -1
	LPWSTR	pstrWC,		//@parm WideChar string to receive converted chars
	int		cchWC,		//@parm Max count for pstrWC or 0 to get cch needed
	LPBOOL 	pfNoCodePage) //@parm Out parm to receive whether code page is on system
{
	BOOL	fNoCodePage = FALSE;			// Default code page is on OS
	int		cch = -1;

	if(CodePage == CP_UTF8)
	{
		DWORD ch,ch1;

		for(cch = 0; cchMB--; )
		{
			ch = ch1 = *(BYTE *)pstrMB++;
			Assert(ch < 256);
			if(ch > 127 && cchMB && IN_RANGE(0x80, *(BYTE *)pstrMB, 0xBF))
			{
				// Need at least 2 bytes of form 110bbbbb 10bbbbbb
				ch1 = ((ch1 & 0x1F) << 6) + (*pstrMB++ & 0x3F);
				cchMB--;
				if(ch >= 0xE0 && cchMB && IN_RANGE(0x80, *(BYTE *)pstrMB, 0xBF))
				{
					// Need at least 3 bytes of form 1110bbbb 10bbbbbb 10bbbbbb
					ch1 = (ch1 << 6) + (*pstrMB++ & 0x3F);
					cchMB--;
					if (ch >= 0xF0 && cchMB && IN_RANGE(0x80, *(BYTE *)pstrMB, 0xBF))
					{
						// Handle 4-byte form for 16 UTF-16 planes above the
						// BMP) expect: 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb
						ch1 = ((ch1 & 0x7FFF) << 6) + (*(BYTE *)pstrMB++ & 0x3F)
							- 0x10000;			// Subtract offset for BMP
						if(ch1 <= 0xFFFFF)		// Fits in 20 bits
						{
							cch++;				// Two 16-bit surrogate codes
							if(cch < cchWC)
								*pstrWC++ = UTF16_LEAD + (ch1 >> 10);
							ch1 = (ch1 & 0x3FF) + UTF16_TRAIL;
							cchMB--;
						}
						else ch1 = '?';
					}
				}
			}
			cch++;
			if(cch < cchWC)
				*pstrWC++ = ch1;
			if(!ch)
				break;
		}
	}
	else if(CodePage != CP_SYMBOL)			// Not SYMBOL_CHARSET
	{
		fNoCodePage = TRUE;					// Default codepage isn't on OS
		if(CodePage >= 0)					// Might be..
		{
			cch = MultiByteToWideChar(
				CodePage, dwFlags, pstrMB, cchMB, pstrWC, cchWC);
			if(cch > 0)
				fNoCodePage = FALSE;		// Codepage is on OS
		}
	}
	if(pfNoCodePage)
		*pfNoCodePage = fNoCodePage;

	if(cch <= 0)
	{			
		// SYMBOL_CHARSET or conversion failed: bytes -> words with
		//  high bytes of 0.  Return count for full conversion

		if(cchWC <= 0)					
			return cchMB >= 0 ? cchMB : (strlen(pstrMB) + 1);

		int cchMBMax = cchMB;

		if(cchMB < 0)					// If negative, use NULL termination
			cchMBMax = tomForward;			//  of pstrMB

		cchMBMax = min(cchMBMax, cchWC);

		for(cch = 0; (cchMB < 0 ? *pstrMB : 1) && cch < cchMBMax; cch++)
		{
			*pstrWC++ = (unsigned char)*pstrMB++;
		}
		
		// NULL-terminate the WC string if the MB string was NULL-terminated,
		// and if there is room in the WC buffer.
		if(cchMB < 0 && cch < cchWC)
		{
			*pstrWC = 0;
			cch++;
		}
	}
	return cch;
}

/*
 *	CW32System::WCTMB (CodePage, dwFlags, pstrWC, cchWC, pstrMB, cchMB,
 *					   pchDefault, pfUsedDef, pfNoCodePage, fTestCodePage)
 *
 *	@mfunc
 *		Convert WideChar (WC) string pstrWC of length cchWC to MultiByte (MB)
 *		string pstrMB of length cchMB according to the flags dwFlags and code
 *		page CodePage.  If CodePage = SYMBOL_CODEPAGE
 *		(usually for SYMBOL_CHARSET strings),
 *		convert each character in pstrWC to a byte, discarding the high byte.
 *		Else call the Win32 WideCharToMultiByte() function.
 *
 *	@rdesc
 *		Count of bytes stored in target string pstrMB
 */
int CW32System::WCTMB(
	INT		CodePage,	//@parm Code page to use for conversion
	DWORD	dwFlags,	//@parm Flags to guide conversion
	LPCWSTR	pstrWC,		//@parm WideChar string to convert
	int		cchWC,		//@parm Count for pstrWC or -1 to use NULL termination
	LPSTR	pstrMB,		//@parm MultiByte string to receive converted chars
	int		cchMB,		//@parm Count of chars (bytes) in pstrMB or 0
	LPCSTR	pchDefault,	//@parm Default char to use if conversion fails
	LPBOOL	pfUsedDef,	//@parm Out parm to receive whether default char used
	LPBOOL 	pfNoCodePage, //@parm Out parm to receive whether code page is on system
	BOOL	fTestCodePage)//@parm Test CodePage could handle the pstrWC
{
	int		cch = -1;						// No chars converted yet
	BOOL	fNoCodePage = FALSE;			// Default code page is on OS

	if(pfUsedDef)							// Default that all chars can be
		*pfUsedDef = FALSE;					//  converted

#ifndef WC_NO_BEST_FIT_CHARS
#define WC_NO_BEST_FIT_CHARS	0x400
#endif

	if (_dwPlatformId == VER_PLATFORM_WIN32_NT &&
		_dwMajorVersion > 4 && !dwFlags)
	{
		dwFlags = WC_NO_BEST_FIT_CHARS;
	}

	if(CodePage == CP_UTF8)					// Convert to UTF8 since OS
	{										// doesn't (pre NT 5.0)
		unsigned ch;
		cch = 0;							// No converted bytes yet
		while(cchWC--)
		{
			ch = *pstrWC++;					// Get Unicode char
			if(ch <= 127)					// It's ASCII
			{
				cch++;
				if(cch < cchMB)
					*pstrMB++ = ch;			// One more converted byte
				if(!ch)						// Quit on NULL termination
					break;
				continue;
			}
			if(ch <= 0x7FF)					// Need 2 bytes of form:
			{								//  110bbbbb 10bbbbbb
				cch += 2;
				if(cch < cchMB)				// Store lead byte
					*pstrMB++ = 0xC0 + (ch >> 6);
			}
			else if(IN_RANGE(UTF16_LEAD, ch, 0xDBFF))
			{								// Unicode surrogate pair
				cch += 4;					// Need 4 bytes of form:
				if(cch < cchMB)				//  11110bbb 10bbbbbb 10bbbbbb
				{							//  10bbbbbb
					AssertSz(IN_RANGE(UTF16_TRAIL, *pstrWC, 0xDFFF),
						"CW32System::WCTMB: illegal surrogate pair");
					cchWC--;
					ch = ((ch & 0x3FF) << 10) + (*pstrWC++ & 0x3FF) + 0x10000;
					*pstrMB++ = 0xF0 + (ch >> 18);
					*pstrMB++ = 0x80 + (ch >> 12 & 0x3F);
					*pstrMB++ = 0x80 + (ch >> 6  & 0x3F);
				}
			}
			else							// Need 3 bytes of form:
			{								//  1110bbbb 10bbbbbb
				cch += 3;					//  10bbbbbb
				if(cch < cchMB)				// Store lead byte followed by
				{							//  first trail byte
					*pstrMB++ = 0xE0 + (ch >> 12);
					*pstrMB++ = 0x80 + (ch >> 6 & 0x3F);
				}
			}
			if(cch < cchMB)					// Store final UTF-8 byte
				*pstrMB++ = 0x80 + (ch & 0x3F);
		}
	}
	else if(CodePage != CP_SYMBOL)
	{
		fNoCodePage = TRUE;					// Default codepage not on OS
		if(CodePage >= 0)					// Might be...
		{
			cch = WideCharToMultiByte(CodePage, dwFlags,
					pstrWC, cchWC, pstrMB, cchMB, pchDefault, pfUsedDef);
			if(cch > 0)
				fNoCodePage = FALSE;		// Found codepage on system
		}
	}
	if(pfNoCodePage)
		*pfNoCodePage = fNoCodePage;

	// Early exit if we are just testing for CodePage
	if (fTestCodePage)
		return cch;

	// SYMBOL_CHARSET, fIsDBCS or conversion failed: low bytes of words ->
	//  bytes
	if(cch <= 0)
	{									
		// Return multibyte count for full conversion. cchWC is correct for
		// single-byte charsets like the 125x's
		if(cchMB <= 0)
		{
			return cchWC >= 0 ? cchWC : wcslen(pstrWC);
		}

		char chDefault = 0;
		BOOL fUseDefaultChar = (pfUsedDef || pchDefault) && CodePage != CP_SYMBOL;

		if(fUseDefaultChar)
		{
			// determine a default char for our home-grown conversion
			if(pchDefault)
			{
				chDefault = *pchDefault;
			}
			else
			{
				static char chSysDef = 0;
				static BOOL fGotSysDef = FALSE;

				// 0x2022 is a math symbol with no conversion to ANSI
				const WCHAR szCantConvert[] = { 0x2022 };
				BOOL fUsedDef;

				if(!fGotSysDef)
				{
					fGotSysDef = TRUE;

					if(!(WideCharToMultiByte
							(CP_ACP, 0, szCantConvert, 1, &chSysDef, 1, NULL,
										&fUsedDef) == 1 && fUsedDef))
					{
						AssertSz(0, "WCTMB():  Unable to determine what the "
									"system uses as its default replacement "
									"character.");
						chSysDef = '?';
					}
				}
				chDefault = chSysDef;
			}
		}

		int cchWCMax = cchWC;

		// If negative, use NULL termination of pstrMB
		if(cchWC < 0)
		{
			cchWCMax = tomForward;
		}

		cchWCMax = min(cchWCMax, cchMB);

		for(cch = 0; (cchWC < 0 ? *pstrWC : 1) && cch < cchWCMax; cch++)
		{
			// TODO(BradO):  Should this be 0x7F in some conversion cases?
			if(fUseDefaultChar && *pstrWC > 0xFF)
			{
				if(pfUsedDef)
				{
					*pfUsedDef = TRUE;
				}
				*pstrMB = chDefault;
			}
			else
			{
				*pstrMB = (BYTE)*pstrWC;
			}
			pstrMB++;
			pstrWC++;
		}

		if(cchWC < 0 && cch < cchMB)
		{
			*pstrMB = 0;
			cch++;
		}
	}
	return cch;
}
/*
 *	CW32System::VerifyFEString(cpg, pstrWC, cchWC, fTestInputCpg)
 *
 *	@mfunc
 *		Verify if the input cpg can handle the pstrWC.
 *		If not, select another FE cpg.
 *
 *	@rdesc
 *		New CodePage for the pstrWC
 */
#define NUMBER_OF_CHARS	64
	
int CW32System::VerifyFEString(
	INT		cpg,			//@parm cpg to format the pstrWC
	LPCWSTR	pstrWC,			//@parm WideChar string to test
	int		cchWC,			//@parm Count for pstrWC
	BOOL	fTestInputCpg)	//@parm test the input cpg only
{
	if (cchWC <=0)
		return cpg;

	int				cpgNew = cpg;
	BOOL			fUsedDef;
	int				cchMB = cchWC * sizeof(WCHAR);
	CTempCharBuf	tcb;
	char			*pstrMB = tcb.GetBuf(cchMB);
	CTempWcharBuf	twcb;
	WCHAR			*pstrWchar = twcb.GetBuf(cchWC);						
	static			int	aiCpg[4] =
		{ CP_JAPAN, CP_KOREAN, CP_CHINESE_TRAD, CP_CHINESE_SIM };	

	if (pstrMB)
	{
		int	cchConverted = WCTMB(cpg, 0, pstrWC, cchWC, pstrMB, cchMB, NULL,
				&fUsedDef, NULL, TRUE);

		if (cchConverted > 0 && !fUsedDef && IsFEFontInSystem(cpg))	
		{
			cchConverted = MBTWC(cpg, 0, pstrMB, cchConverted, pstrWchar, cchWC, NULL);

			if (cchConverted == cchWC && memcmp(pstrWC, pstrWchar, cchMB) == 0)			
				goto Exit;					// Found it
		}		
		
		if (fTestInputCpg)				// Only need to test the input cpg			
			cpgNew = -1;				// Indicate cpg doesn't support the string
		else
		{
			// If no conversion or if the default character is used or
			// no such FE font in system,
			// it means that this cpg may not be the right choice.
			// Let's try other FE cpg.
			for (int i=0; i < 4; i++)
			{
				if (cpg != aiCpg[i])
				{
					cchConverted = WCTMB(aiCpg[i], 0, pstrWC, cchWC, pstrMB, cchMB, NULL,
						&fUsedDef, NULL, TRUE);

					if (cchConverted > 0 && !fUsedDef && IsFEFontInSystem(aiCpg[i]))	
					{
						cchConverted = MBTWC(aiCpg[i], 0, pstrMB, cchConverted, pstrWchar, cchWC, NULL);

						if (cchConverted == cchWC && memcmp(pstrWC, pstrWchar, cchMB) == 0)	
						{
							cpgNew = aiCpg[i];	// Found it
							break;
						}
					}
				}
			}
		}
	}			

Exit:

	return cpgNew;
}

int __cdecl CW32System::sprintf(char * buff, char *fmt, ...)
{
	va_list	marker;

	va_start(marker, fmt);
	int cb = W32->WvsprintfA(0x07FFFFFFF, buff, fmt, marker);
	va_end(marker);

	return cb;
}

//NOTE: Keep this at the end because we want any allocations in w32sys.cpp to be validated
//just like the rest of RichEdit.
///////////////////////////////  Memory mamagement  /////////////////////////////////

#ifdef DEBUG
#undef PvAlloc
#undef PvReAlloc
#undef FreePv
#undef new

MST vrgmst[100];

typedef struct tagPVH //PV Header
{
	char	*szFile;
	int		line;
	tagPVH	*ppvhNext;
	int		cbAlloc;	//On Win'95, the size returned is not the size allocated.
	int		magicPvh;	//Should be last
} PVH;
#define cbPvh (sizeof(PVH))

typedef struct //PV Tail
{
	int		magicPvt; //Must be first
} PVT;

#define cbPvt (sizeof(PVT))
#define cbPvDebug (cbPvh + cbPvt)

void *vpHead = 0;

/*
 *	CW32System::UpdateMst(void)
 *
 *	@mfunc Fills up the vrgmst structure with summary information about our memory
 *	usage.
 *
 *	@rdesc
 *		void
 */
void UpdateMst(void)
{
	W32->ZeroMemory(vrgmst, sizeof(vrgmst));

	PVH		*ppvh;
	MST		*pmst;

	ppvh = (PVH*) vpHead;

	while (ppvh != 0)
	{
		pmst = vrgmst;

		//Look for entry in list...
		while (pmst->szFile)
		{
			if (strcmp(pmst->szFile, ppvh->szFile) == 0)
			{
				pmst->cbAlloc += ppvh->cbAlloc;
				break;
			}
			pmst++;
		}

		if (pmst->szFile == 0)
		{
			pmst->szFile = ppvh->szFile;
			pmst->cbAlloc = ppvh->cbAlloc;
		}

		ppvh = ppvh->ppvhNext;
	}
}

/*
 *	PvDebugValidate(void)
 *
 *	@func Verifies the the node is proper.  Pass in a pointer to the users data
 *	(after the header node.)
 *
 *	@rdesc
 *		void
 */
void PvDebugValidate(void *pv)
{
	PVH	*ppvh;
	PVT *ppvt;

	ppvh = (PVH*) ((char*) pv - cbPvh);
	ppvt = (PVT*) ((char*) pv + ppvh->cbAlloc);

	AssertSz(ppvh->magicPvh == 0x12345678, "PvDebugValidate: header bytes are corrupt");
	AssertSz(ppvt->magicPvt == 0xfedcba98, "PvDebugValidate: tail bytes are corrupt");
}

/*
 *	CW32System::PvSet(pv, szFile, line)
 *
 *	@mfunc Sets a different module and line number for
 *
 *	@rdesc
 *		void
 */
void CW32System::PvSet(void *pv, char *szFile, int line)
{
	if (pv == 0)
		return;

	PvDebugValidate(pv);
	PVH *ppvh = (PVH*) ((char*) pv - cbPvh);

	ppvh->szFile = szFile;
	ppvh->line = line;
}
/*
 *	CW32System::PvAllocDebug(cb, uiMemFlags, szFile, line)
 *
 *	@mfunc Allocates a generic (void*) pointer. This is a debug only routine which
 *	tracks the allocation.
 *
 *	@rdesc
 *		void
 */
void* CW32System::PvAllocDebug(ULONG cb, UINT uiMemFlags, char *szFile, int line)
{
	void	*pv;

	pv = PvAlloc(cb + cbPvDebug, uiMemFlags);
	if (!pv)
		return 0;

	PVH	*ppvh;
	PVT *ppvt;

	ppvt = (PVT*) ((char*) pv + cb + cbPvh);
	ppvh = (PVH*) pv;

	ZeroMemory(ppvh, sizeof(PVH));
	ppvh->magicPvh = 0x12345678;
	ppvt->magicPvt = 0xfedcba98;
	ppvh->szFile = szFile;
	ppvh->line = line;
	ppvh->cbAlloc = cb;

	ppvh->ppvhNext = (PVH*) vpHead;
	vpHead = pv;

	return (char*) pv + cbPvh;
}

/*
 *	CW32System::PvReAllocDebug(pv, cb, szFile, line)
 *
 *	@mfunc ReAllocates a generic (void*) pointer. This is a debug only routine which
 *	tracks the allocation.
 *
 *	@rdesc
 *		void
 */
void* CW32System::PvReAllocDebug(void *pv, ULONG cb, char *szFile, int line)
{
	void	*pvNew;
	PVH	*ppvh, *ppvhHead, *ppvhTail;
	PVT *ppvt;
	ppvh = (PVH*) ((char*) pv - cbPvh);

	if (!pv)
		return PvAllocDebug(cb, 0, szFile, line);

	PvDebugValidate(pv);

	pvNew = PvReAlloc((char*) pv - cbPvh, cb + cbPvDebug);

	if (!pvNew)
		return 0;

	ppvt = (PVT*) ((char*) pvNew + cb + cbPvh);
	ppvh = (PVH*) pvNew;
	ppvh->cbAlloc = cb;

	//Put the new trailer bytes in.
	ppvt->magicPvt = 0xfedcba98;

	//Make the pointer list up to date again
	if (pv != pvNew)
	{
		ppvhTail = 0;
		ppvhHead = (PVH*) vpHead;

		while ((char*)ppvhHead != (char*)pv - cbPvh)
		{
			AssertSz(ppvhHead, "entry not found in list.");
			ppvhTail = ppvhHead;
			ppvhHead = (PVH*) ppvhHead->ppvhNext;
		}

		if (ppvhTail == 0)
			vpHead = pvNew;
		else
			ppvhTail->ppvhNext = (PVH*) pvNew;
	}

	return (char*) pvNew + cbPvh;
}

/*
 *	CW32System::FreePvDebug(pv)
 *
 *	@mfunc Returns a pointer when you are done with it.
 *
 *	@rdesc
 *		void
 */
void CW32System::FreePvDebug(void *pv)
{
	if (!pv)
		return;

	PvDebugValidate(pv);

	PVH	*ppvhHead, *ppvhTail, *ppvh;

	AssertSz(vpHead, "Deleting from empty free list.");

	ppvh = (PVH*) ((char*) pv - cbPvh);
	
	//Search and remove the entry from the list
	ppvhTail = 0;
	ppvhHead = (PVH*) vpHead;

	while ((char*) ppvhHead != ((char*) pv - cbPvh))
	{
		AssertSz(ppvhHead, "entry not found in list.");
		ppvhTail = ppvhHead;
		ppvhHead = (PVH*) ppvhHead->ppvhNext;
	}

	if (ppvhTail == 0)
		vpHead = ppvhHead->ppvhNext;
	else
		ppvhTail->ppvhNext = ppvhHead->ppvhNext;

	FreePv((char*) pv - cbPvh);
}

/*
 *	CatchLeaks(void)
 *
 *	@func Displays any memory leaks in a dialog box.
 *
 *	@rdesc
 *		void
 */
void CatchLeaks(void)
{
	PVH		*ppvh;
	char szLeak[512];

	ppvh = (PVH*) vpHead;
	while (ppvh != 0)
	{
		wsprintfA(szLeak, "Memory Leak of %d bytes: -- File: %s, Line: %d", ppvh->cbAlloc, ppvh->szFile, ppvh->line);
	    if (NULL != pfnAssert)
		{
			// if we have an assert hook, give the user a chance to process the leak message
			if (pfnAssert(szLeak, ppvh->szFile, &ppvh->line))
			{
				// hook returned true, show the message box
				MessageBoxA(NULL, szLeak, "", MB_OK);				
			}
		}
		else
		{
			MessageBoxA(NULL, szLeak, "", MB_OK);
		}
		ppvh = ppvh->ppvhNext;
	}
}

void* _cdecl operator new (size_t size, char *szFile, int line)
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "new");

	return W32->PvAllocDebug(size, GMEM_ZEROINIT, szFile, line);
}

void _cdecl operator delete (void* pv)
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "delete");

	W32->FreePvDebug(pv);
}

#else //DEBUG

void* _cdecl operator new (size_t size)
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "new");

	return W32->PvAlloc(size, GMEM_ZEROINIT);
}

void _cdecl operator delete (void* pv)
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "delete");

	W32->FreePv(pv);
}

#endif //DEBUG


/*
 *	PvAlloc (cbBuf, uiMemFlags)
 *
 *	@mfunc	memory allocation.  Similar to GlobalAlloc.
 *
 *	@comm	The only flag of interest is GMEM_ZEROINIT, which
 *			specifies that memory should be zeroed after allocation.
 */
PVOID CW32System::PvAlloc(
	ULONG	cbBuf, 			//@parm	Count of bytes to allocate
	UINT	uiMemFlags)		//@parm Flags controlling allocation
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "PvAlloc");
	
	void * pv = LocalAlloc(LMEM_FIXED, cbBuf);

	if( pv && (uiMemFlags & GMEM_ZEROINIT) )
		ZeroMemory(pv, cbBuf);
	
	return pv;
}

/*
 *	PvReAlloc	(pvBuf, cbBuf)
 *
 *	@mfunc	memory reallocation.
 *
 */
PVOID CW32System::PvReAlloc(
	PVOID	pvBuf, 		//@parm Buffer to reallocate
	DWORD	cbBuf)		//@parm New size of buffer
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "PvReAlloc");

	if(pvBuf)
		return LocalReAlloc(pvBuf, cbBuf, LMEM_MOVEABLE);

	return LocalAlloc(LMEM_FIXED, cbBuf);
}

/*
 *	FreePv (pvBuf)
 *
 *	@mfunc	frees memory
 *
 *	@rdesc	void
 */
void CW32System::FreePv(
	PVOID pvBuf)		//@parm Buffer to free
{
	TRACEBEGIN(TRCSUBSYSEDIT, TRCSCOPEINTERN, "FreePv");

	if(pvBuf)
		LocalFree(pvBuf);
}


/*
 *	Mirroring API (only in BiDi Win98 and NT5 upward)
 *
 *	@mfunc	Get/Set DC mirroring effect
 *
 */

DWORD WINAPI GetLayoutStub(HDC hdc)
{
	return 0;
}

DWORD WINAPI SetLayoutStub(HDC hdc, DWORD dwLayout)
{
	return 0;
}

DWORD WINAPI GetLayoutInit(HDC hdc)
{
	CLock		lock;
	HINSTANCE	hMod = ::GetModuleHandleA("GDI32.DLL");
	Assert(hMod);

	W32->_pfnGetLayout = (PFN_GETLAYOUT)GetProcAddress(hMod, "GetLayout");

	if (!W32->_pfnGetLayout)
		W32->_pfnGetLayout = &GetLayoutStub;

	return W32->_pfnGetLayout(hdc);
}

DWORD WINAPI SetLayoutInit(HDC hdc, DWORD dwLayout)
{
	CLock		lock;
	HINSTANCE	hMod = ::GetModuleHandleA("GDI32.DLL");
	Assert(hMod);

	W32->_pfnSetLayout = (PFN_SETLAYOUT)GetProcAddress(hMod, "SetLayout");

	if (!W32->_pfnSetLayout)
		W32->_pfnSetLayout = &SetLayoutStub;

	return W32->_pfnSetLayout(hdc, dwLayout);
}

PFN_GETLAYOUT	CW32System::_pfnGetLayout = &GetLayoutInit;
PFN_SETLAYOUT	CW32System::_pfnSetLayout = &SetLayoutInit;