checkpoint. adding support for unicode in clipboard.

base/CUnicode.cpp

@@ -0,0 +1,530 @@
+#include "CUnicode.h"
+#include <string.h>
+
+//
+// local utility functions
+//
+
+inline
+static
+UInt16
+decode16(const UInt8* n)
+{
+	union x16 {
+		UInt8	n8[2];
+		UInt16	n16;
+	} c;
+	c.n8[0] = n[0];
+	c.n8[1] = n[1];
+	return c.n16;
+}
+
+inline
+static
+UInt32
+decode32(const UInt8* n)
+{
+	union x32 {
+		UInt8	n8[4];
+		UInt32	n32;
+	} c;
+	c.n8[0] = n[0];
+	c.n8[1] = n[1];
+	c.n8[2] = n[2];
+	c.n8[3] = n[3];
+	return c.n32;
+}
+
+//
+// CUnicode
+//
+
+UInt32					CUnicode::s_invalid = 0x0000ffff;
+
+CString
+CUnicode::UTF8ToUCS2(const CString& src)
+{
+	// get size of input string and reserve some space in output.
+	// include UTF8's nul terminator.
+	UInt32 n = src.size() + 1;
+	CString dst;
+	dst.reserve(2 * n);
+
+	// convert each character
+	const UInt8* data = reinterpret_cast<const UInt8*>(src.c_str());
+	while (n > 0) {
+		UInt32 c = fromUTF8(data, n);
+		if (c != s_invalid && c < 0x0000ffff) {
+			UInt16 ucs2 = static_cast<UInt16>(c);
+			dst.append(reinterpret_cast<const char*>(&ucs2), 2);
+		}
+	}
+
+	return dst;
+}
+
+CString
+CUnicode::UTF8ToUCS4(const CString& src)
+{
+	// get size of input string and reserve some space in output.
+	// include UTF8's nul terminator.
+	UInt32 n = src.size() + 1;
+	CString dst;
+	dst.reserve(4 * n);
+
+	// convert each character
+	const UInt8* data = reinterpret_cast<const UInt8*>(src.c_str());
+	while (n > 0) {
+		UInt32 c = fromUTF8(data, n);
+		if (c != s_invalid) {
+			dst.append(reinterpret_cast<const char*>(&c), 4);
+		}
+	}
+
+	return dst;
+}
+
+CString
+CUnicode::UTF8ToUTF16(const CString& src)
+{
+	// get size of input string and reserve some space in output.
+	// include UTF8's nul terminator.
+	UInt32 n = src.size() + 1;
+	CString dst;
+	dst.reserve(2 * n);
+
+	// convert each character
+	const UInt8* data = reinterpret_cast<const UInt8*>(src.c_str());
+	while (n > 0) {
+		UInt32 c = fromUTF8(data, n);
+		if (c != s_invalid && c < 0x0010ffff) {
+			if (c < 0x00010000) {
+				UInt16 ucs2 = static_cast<UInt16>(c);
+				dst.append(reinterpret_cast<const char*>(&ucs2), 2);
+			}
+			else {
+				c -= 0x00010000;
+				UInt16 utf16h = static_cast<UInt16>(c >> 10) + 0xd800;
+				UInt16 utf16l = (static_cast<UInt16>(c) & 0x03ff) + 0xdc00;
+				dst.append(reinterpret_cast<const char*>(&utf16h), 2);
+				dst.append(reinterpret_cast<const char*>(&utf16l), 2);
+			}
+		}
+	}
+
+	return dst;
+}
+
+CString
+CUnicode::UTF8ToUTF32(const CString& src)
+{
+	// FIXME -- should ensure dst has no characters over U-0010FFFF
+	return UTF8ToUCS4(src);
+}
+
+CString
+CUnicode::UCS2ToUTF8(const CString& src)
+{
+	UInt32 n = src.size() >> 1;
+	return doUCS2ToUTF8(reinterpret_cast<const UInt8*>(src.data()), n);
+}
+
+CString
+CUnicode::UCS4ToUTF8(const CString& src)
+{
+	UInt32 n = src.size() >> 2;
+	return doUCS4ToUTF8(reinterpret_cast<const UInt8*>(src.data()), n);
+}
+
+CString
+CUnicode::UTF16ToUTF8(const CString& src)
+{
+	UInt32 n = src.size() >> 1;
+	return doUTF16ToUTF8(reinterpret_cast<const UInt8*>(src.data()), n);
+}
+
+CString
+CUnicode::UTF32ToUTF8(const CString& src)
+{
+	UInt32 n = src.size() >> 2;
+	return doUTF32ToUTF8(reinterpret_cast<const UInt8*>(src.data()), n);
+}
+
+CString
+CUnicode::UTF8ToText(const CString& src)
+{
+	// convert to wide char
+	wchar_t* tmp = UTF8ToWideChar(src);
+
+	// get length of multibyte string
+	mbstate_t state;
+	memset(&state, 0, sizeof(state));
+	const wchar_t* scratch = tmp;
+	size_t len = wcsrtombs(NULL, &scratch, 0, &state);
+	if (len == (size_t)-1) {
+		// invalid character in src
+		delete[] tmp;
+		return CString();
+	}
+
+	// convert to multibyte
+	scratch = tmp;
+	char* dst = new char[len + 1];
+	wcsrtombs(dst, &scratch, len + 1, &state);
+	CString text(dst);
+
+	// clean up
+	delete[] dst;
+	delete[] tmp;
+
+	return text;
+}
+
+CString
+CUnicode::textToUTF8(const CString& src)
+{
+	// get length of wide char string
+	mbstate_t state;
+	memset(&state, 0, sizeof(state));
+	const char* scratch = src.c_str();
+	size_t len = mbsrtowcs(NULL, &scratch, 0, &state);
+	if (len == (size_t)-1) {
+		// invalid character in src
+		return CString();
+	}
+
+	// convert multibyte to wide char
+	scratch = src.c_str();
+	wchar_t* dst = new wchar_t[len + 1];
+	mbsrtowcs(dst, &scratch, len + 1, &state);
+
+	// convert to UTF8
+	CString utf8 = wideCharToUTF8(dst);
+
+	// clean up
+	delete[] dst;
+
+	return utf8;
+}
+
+wchar_t*
+CUnicode::UTF8ToWideChar(const CString& src)
+{
+	// convert to platform's wide character encoding.
+	// note -- this must include a wide nul character (independent of
+	// the CString's nul character).
+#if WINDOWS_LIKE
+	CString tmp = UTF8ToUCS16(src);
+	UInt32 size = tmp.size() >> 1;
+#elif UNIX_LIKE
+	CString tmp = UTF8ToUCS4(src);
+	UInt32 size = tmp.size() >> 2;
+#endif
+
+	// copy to a wchar_t array
+	wchar_t* dst = new wchar_t[size];
+	::memcpy(dst, src.data(), sizeof(wchar_t) * size);
+	return dst;
+}
+
+CString
+CUnicode::wideCharToUTF8(const wchar_t* src)
+{
+	// convert from platform's wide character encoding.
+	// note -- this must include a wide nul character (independent of
+	// the CString's nul character).
+#if WINDOWS_LIKE
+	return doUCS16ToUTF8(reinterpret_cast<const UInt8*>(src), wcslen(src));
+#elif UNIX_LIKE
+	return doUCS4ToUTF8(reinterpret_cast<const UInt8*>(src), wcslen(src));
+#endif
+}
+
+CString
+CUnicode::doUCS2ToUTF8(const UInt8* data, UInt32 n)
+{
+	// make some space
+	CString dst;
+	dst.reserve(n);
+
+	// convert each character
+	for (; n > 0; data += 2, --n) {
+		UInt32 c = decode16(data);
+		toUTF8(dst, c);
+	}
+
+	// remove extra trailing nul
+	if (dst.size() > 0 && dst[dst.size() - 1] == '\0') {
+		dst.resize(dst.size() - 1);
+	}
+
+	return dst;
+}
+
+CString
+CUnicode::doUCS4ToUTF8(const UInt8* data, UInt32 n)
+{
+	// make some space
+	CString dst;
+	dst.reserve(n);
+
+	// convert each character
+	for (; n > 0; data += 4, --n) {
+		UInt32 c = decode32(data);
+		toUTF8(dst, c);
+	}
+
+	// remove extra trailing nul
+	if (dst.size() > 0 && dst[dst.size() - 1] == '\0') {
+		dst.resize(dst.size() - 1);
+	}
+
+	return dst;
+}
+
+CString
+CUnicode::doUTF16ToUTF8(const UInt8* data, UInt32 n)
+{
+	// make some space
+	CString dst;
+	dst.reserve(n);
+
+	// convert each character
+	for (; n > 0; data += 2, --n) {
+		UInt32 c = decode16(data);
+		if (c < 0x0000d800 || c > 0x0000dfff) {
+			toUTF8(dst, c);
+		}
+		else if (n == 1) {
+			// error -- missing second word
+		}
+		else if (c >= 0x0000d800 && c <= 0x0000dbff) {
+			UInt32 c2 = decode16(data);
+			data += 2;
+			--n;
+			if (c2 < 0x0000dc00 || c2 > 0x0000dfff) {
+				// error -- [d800,dbff] not followed by [dc00,dfff]
+			}
+			else {
+				c = (((c - 0x0000d800) << 10) | (c2 - 0x0000dc00)) + 0x00010000;
+				toUTF8(dst, c);
+			}
+		}
+		else {
+			// error -- [dc00,dfff] without leading [d800,dbff]
+		}
+	}
+
+	// remove extra trailing nul
+	if (dst.size() > 0 && dst[dst.size() - 1] == '\0') {
+		dst.resize(dst.size() - 1);
+	}
+
+	return dst;
+}
+
+CString
+CUnicode::doUTF32ToUTF8(const UInt8* data, UInt32 n)
+{
+	// FIXME -- should check that src has no characters over U-0010FFFF
+	return doUCS4ToUTF8(data, n);
+}
+
+UInt32
+CUnicode::fromUTF8(const UInt8*& data, UInt32& n)
+{
+	assert(data != NULL);
+	assert(n    != 0);
+
+	// compute character encoding length, checking for overlong
+	// sequences (i.e. characters that don't use the shortest
+	// possible encoding).
+	UInt32 size;
+	if (data[0] < 0x80) {
+		// 0xxxxxxx
+		size = 1;
+	}
+	else if (data[0] < 0xc0) {
+		// 10xxxxxx -- in the middle of a multibyte character.  skip
+		// until we find a start byte and return error.
+		do {
+			--n;
+			++data;
+		} while (n > 0 && (data[0] & 0xc0) == 0x80);
+		return s_invalid;
+	}
+	else if (data[0] < 0xe0) {
+		// 110xxxxx
+		size = 2;
+	}
+	else if (data[0] < 0xf0) {
+		// 1110xxxx
+		size = 3;
+	}
+	else if (data[0] < 0xf8) {
+		// 11110xxx
+		size = 4;
+	}
+	else if (data[0] < 0xfc) {
+		// 111110xx
+		size = 5;
+	}
+	else if (data[0] < 0xfe) {
+		// 1111110x
+		size = 6;
+	}
+	else {
+		// invalid sequence.  dunno how many bytes to skip so skip one.
+		--n;
+		++data;
+		return s_invalid;
+	}
+
+	// make sure we have enough data
+	if (size > n) {
+		data += n;
+		n     = 0;
+		return s_invalid;
+	}
+
+	// extract character
+	UInt32 c;
+	switch (size) {
+	case 1:
+		c = static_cast<UInt32>(data[0]);
+		break;
+
+	case 2:
+		c = ((static_cast<UInt32>(data[0]) & 0x1f) <<  6) |
+			((static_cast<UInt32>(data[1]) & 0x3f)      );
+		break;
+
+	case 3:
+		c = ((static_cast<UInt32>(data[0]) & 0x0f) << 12) |
+			((static_cast<UInt32>(data[1]) & 0x3f) <<  6) |
+			((static_cast<UInt32>(data[2]) & 0x3f)      );
+		break;
+
+	case 4:
+		c = ((static_cast<UInt32>(data[0]) & 0x07) << 18) |
+			((static_cast<UInt32>(data[1]) & 0x3f) << 12) |
+			((static_cast<UInt32>(data[1]) & 0x3f) <<  6) |
+			((static_cast<UInt32>(data[1]) & 0x3f)      );
+		break;
+
+	case 5:
+		c = ((static_cast<UInt32>(data[0]) & 0x03) << 24) |
+			((static_cast<UInt32>(data[1]) & 0x3f) << 18) |
+			((static_cast<UInt32>(data[1]) & 0x3f) << 12) |
+			((static_cast<UInt32>(data[1]) & 0x3f) <<  6) |
+			((static_cast<UInt32>(data[1]) & 0x3f)      );
+		break;
+
+	case 6:
+		c = ((static_cast<UInt32>(data[0]) & 0x01) << 30) |
+			((static_cast<UInt32>(data[1]) & 0x3f) << 24) |
+			((static_cast<UInt32>(data[1]) & 0x3f) << 18) |
+			((static_cast<UInt32>(data[1]) & 0x3f) << 12) |
+			((static_cast<UInt32>(data[1]) & 0x3f) <<  6) |
+			((static_cast<UInt32>(data[1]) & 0x3f)      );
+		break;
+
+	default:
+		assert(0 && "invalid size");
+	}
+
+	// update parameters
+	data += size;
+	n    -= size;
+
+	// check for characters that didn't use the smallest possible encoding
+	static UInt32 s_minChar[] = {
+		0,
+		0x00000000,
+		0x00000080,
+		0x00000800,
+		0x00010000,
+		0x00200000,
+		0x04000000
+	};
+	if (c < s_minChar[size]) {
+		return s_invalid;
+	}
+
+	// check that all bytes after the first have the pattern 10xxxxxx.
+	UInt8 a = 0x80;
+	switch (size) {
+	case 6:
+		a |= data[5];
+		// fall through
+
+	case 5:
+		a |= data[4];
+		// fall through
+
+	case 4:
+		a |= data[3];
+		// fall through
+
+	case 3:
+		a |= data[2];
+		// fall through
+
+	case 2:
+		a |= data[1];
+	}
+	if ((a & 0xc0) != 0x80) {
+		return s_invalid;
+	}
+
+	return c;
+}
+
+void
+CUnicode::toUTF8(CString& dst, const UInt32 c)
+{
+	UInt8 data[6];
+
+	if (c < 0x00000080) {
+		data[0] = static_cast<UInt8>(c);
+		dst.append(reinterpret_cast<char*>(data), 1);
+	}
+	else if (c < 0x00000800) {
+		data[0] = static_cast<UInt8>((c >>  6) & 0x0000001f) + 0xc0;
+		data[1] = static_cast<UInt8>(c         & 0x0000003f) + 0x80;
+		dst.append(reinterpret_cast<char*>(data), 2);
+	}
+	else if (c < 0x00010000) {
+		data[0] = static_cast<UInt8>((c >> 12) & 0x0000000f) + 0xe0;
+		data[1] = static_cast<UInt8>((c >>  6) & 0x0000003f) + 0x80;
+		data[2] = static_cast<UInt8>(c         & 0x0000003f) + 0x80;
+		dst.append(reinterpret_cast<char*>(data), 3);
+	}
+	else if (c < 0x00200000) {
+		data[0] = static_cast<UInt8>((c >> 18) & 0x00000007) + 0xf0;
+		data[1] = static_cast<UInt8>((c >> 12) & 0x0000003f) + 0x80;
+		data[2] = static_cast<UInt8>((c >>  6) & 0x0000003f) + 0x80;
+		data[3] = static_cast<UInt8>(c         & 0x0000003f) + 0x80;
+		dst.append(reinterpret_cast<char*>(data), 4);
+	}
+	else if (c < 0x04000000) {
+		data[0] = static_cast<UInt8>((c >> 24) & 0x00000003) + 0xf8;
+		data[1] = static_cast<UInt8>((c >> 18) & 0x0000003f) + 0x80;
+		data[2] = static_cast<UInt8>((c >> 12) & 0x0000003f) + 0x80;
+		data[3] = static_cast<UInt8>((c >>  6) & 0x0000003f) + 0x80;
+		data[4] = static_cast<UInt8>(c         & 0x0000003f) + 0x80;
+		dst.append(reinterpret_cast<char*>(data), 5);
+	}
+	else if (c < 0x80000000) {
+		data[0] = static_cast<UInt8>((c >> 30) & 0x00000001) + 0xfc;
+		data[1] = static_cast<UInt8>((c >> 24) & 0x0000003f) + 0x80;
+		data[2] = static_cast<UInt8>((c >> 18) & 0x0000003f) + 0x80;
+		data[3] = static_cast<UInt8>((c >> 12) & 0x0000003f) + 0x80;
+		data[4] = static_cast<UInt8>((c >>  6) & 0x0000003f) + 0x80;
+		data[5] = static_cast<UInt8>(c         & 0x0000003f) + 0x80;
+		dst.append(reinterpret_cast<char*>(data), 6);
+	}
+	else {
+		// invalid character
+	}
+}

base/CUnicode.h

@@ -0,0 +1,48 @@
+#ifndef CUNICODE_H
+#define CUNICODE_H
+
+#include "CString.h"
+#include "BasicTypes.h"
+#include <wchar.h>
+
+class CUnicode {
+public:
+	static CString		UTF8ToUCS2(const CString&);
+	static CString		UTF8ToUCS4(const CString&);
+	static CString		UTF8ToUTF16(const CString&);
+	static CString		UTF8ToUTF32(const CString&);
+
+	static CString		UCS2ToUTF8(const CString&);
+	static CString		UCS4ToUTF8(const CString&);
+	static CString		UTF16ToUTF8(const CString&);
+	static CString		UTF32ToUTF8(const CString&);
+
+	// convert UTF-8 to/from the current locale's encoding
+	static CString		UTF8ToText(const CString&);
+	static CString		textToUTF8(const CString&);
+
+private:
+	// convert UTF8 to nul terminated wchar_t string (using whatever
+	// encoding is native to the platform).  caller must delete[]
+	// the returned string.
+	static wchar_t*		UTF8ToWideChar(const CString&);
+
+	// convert nul terminated wchar_t string (in platform's native
+	// encoding) to UTF8.
+	static CString		wideCharToUTF8(const wchar_t*);
+
+	// internal conversion to UTF8
+	static CString		doUCS2ToUTF8(const UInt8* src, UInt32 n);
+	static CString		doUCS4ToUTF8(const UInt8* src, UInt32 n);
+	static CString		doUTF16ToUTF8(const UInt8* src, UInt32 n);
+	static CString		doUTF32ToUTF8(const UInt8* src, UInt32 n);
+
+	// convert characters to/from UTF8
+	static UInt32		fromUTF8(const UInt8*& src, UInt32& size);
+	static void			toUTF8(CString& dst, const UInt32 c);
+
+private:
+	static UInt32		s_invalid;
+};
+
+#endif

base/Makefile.am

@@ -8,12 +8,14 @@ libbase_a_SOURCES =			\
 	CLog.cpp				\
 	CStopwatch.cpp			\
 	CString.cpp				\
+	CUnicode.cpp			\
 	XBase.cpp				\
 	BasicTypes.h			\
 	CFunctionJob.h			\
 	CLog.h					\
 	CStopwatch.h			\
 	CString.h				\
+	CUnicode.h				\
 	IInterface.h			\
 	IJob.h					\
 	TMethodJob.h			\

platform/CXWindowsClipboard.cpp

@@ -1,4 +1,7 @@
 #include "CXWindowsClipboard.h"
+#include "CXWindowsClipboardTextConverter.h"
+#include "CXWindowsClipboardUCS2Converter.h"
+#include "CXWindowsClipboardUTF8Converter.h"
 #include "CXWindowsUtil.h"
 #include "CThread.h"
 #include "CLog.h"
@@ -52,6 +55,20 @@ CXWindowsClipboard::CXWindowsClipboard(Display* display,
 		break;
 	}
 
+	// add converters
+	m_converters.push_back(new CXWindowsClipboardUTF8Converter(m_display,
+								"text/plain;charset=UTF-8"));
+	m_converters.push_back(new CXWindowsClipboardUTF8Converter(m_display,
+								"UTF8_STRING"));
+	m_converters.push_back(new CXWindowsClipboardUCS2Converter(m_display,
+								"text/plain;charset=ISO-10646-UCS-2"));
+	m_converters.push_back(new CXWindowsClipboardUCS2Converter(m_display,
+								"text/unicode"));
+	m_converters.push_back(new CXWindowsClipboardTextConverter(m_display,
+								"text/plain"));
+	m_converters.push_back(new CXWindowsClipboardTextConverter(m_display,
+								"STRING"));
+
 	// we have no data
 	clearCache();
 }
@@ -59,6 +76,7 @@ CXWindowsClipboard::CXWindowsClipboard(Display* display,
 CXWindowsClipboard::~CXWindowsClipboard()
 {
 	clearReplies();
+	clearConverters();
 }
 
 void
@@ -124,7 +142,7 @@ CXWindowsClipboard::addSimpleRequest(Window requestor,
 
 	// handle targets
 	CString data;
-	Atom type = None;
+	Atom type  = None;
 	int format = 0;
 	if (target == m_atomTargets) {
 		type = getTargetsData(data, &format);
@@ -132,9 +150,17 @@ CXWindowsClipboard::addSimpleRequest(Window requestor,
 	else if (target == m_atomTimestamp) {
 		type = getTimestampData(data, &format);
 	}
-	else if (target == m_atomString ||
+	else {
-			target == m_atomText) {
+		IXWindowsClipboardConverter* converter = getConverter(target);
-		type = getStringData(data, &format);
+		IClipboard::EFormat clipboardFormat = converter->getFormat();
+		if (!m_added[clipboardFormat]) {
+			type = None;
+		}
+		else {
+			type   = converter->getAtom();
+			format = converter->getDataSize();
+			data   = converter->fromIClipboard(m_data[clipboardFormat]);
+		}
 	}
 
 	if (type != None) {
@@ -334,17 +360,41 @@ CXWindowsClipboard::get(EFormat format) const
 	return m_data[format];
 }
 
-IClipboard::EFormat
+void
-CXWindowsClipboard::getFormat(Atom src) const
+CXWindowsClipboard::clearConverters()
 {
-	// FIXME -- handle more formats (especially mime-type-like formats
+	for (ConverterList::iterator index = m_converters.begin();
-	// and various character encodings like unicode).
+								index != m_converters.end(); ++index) {
-	if (src == m_atomString ||
+		delete *index;
-		src == m_atomText /*||
-		src == m_atomCompoundText*/) {
-		return IClipboard::kText;
 	}
-	return IClipboard::kNumFormats;
+	m_converters.clear();
+}
+
+IXWindowsClipboardConverter*
+CXWindowsClipboard::getConverter(Atom target, bool onlyIfNotAdded) const
+{
+	IXWindowsClipboardConverter* converter = NULL;
+	for (ConverterList::const_iterator index = m_converters.begin();
+								index != m_converters.end(); ++index) {
+		converter = *index;
+		if (converter->getAtom() == target) {
+			break;
+		}
+	}
+	if (converter == NULL) {
+		log((CLOG_DEBUG1 "  no converter for target %d", target));
+		return NULL;
+	}
+
+	// optionally skip already handled targets
+	if (onlyIfNotAdded) {
+		if (m_added[converter->getFormat()]) {
+			log((CLOG_DEBUG1 "  skipping handled format %d", converter->getFormat()));
+			return NULL;
+		}
+	}
+
+	return converter;
 }
 
 void
@@ -438,44 +488,29 @@ CXWindowsClipboard::icccmFillCache()
 	const Atom* targets = reinterpret_cast<const Atom*>(data.data());
 	const UInt32 numTargets = data.size() / sizeof(Atom);
 	for (UInt32 i = 0; i < numTargets; ++i) {
-		// determine the expected clipboard format
+		// see if we have a converter for this target
 		Atom target = targets[i];
-		IClipboard::EFormat expectedFormat = getFormat(target);
+		IXWindowsClipboardConverter* converter = getConverter(target, true);
-		if (expectedFormat == IClipboard::kNumFormats) {
+		if (converter == NULL) {
-			log((CLOG_DEBUG1 "  no format for target %d", target));
-			continue;
-		}
-		log((CLOG_DEBUG1 " source target %d -> %d", target, expectedFormat));
-
-		// skip already handled targets
-		if (m_added[expectedFormat]) {
-			log((CLOG_DEBUG1 "  skipping handled format %d", expectedFormat));
 			continue;
 		}
 
+		// get the data
 		Atom actualTarget;
 		CString targetData;
 		if (!icccmGetSelection(target, &actualTarget, &targetData)) {
 			log((CLOG_DEBUG1 "  no data for target", target));
 			continue;
 		}
-		logc(actualTarget != target, (CLOG_DEBUG1 "  actual target is %d", actualTarget));
+		if (actualTarget != target) {
-
+			log((CLOG_DEBUG1 "  expected (%d) and actual (%d) targets differ", target, actualTarget));
-		// use the actual format, not the expected
-		IClipboard::EFormat actualFormat = getFormat(actualTarget);
-		if (actualFormat == IClipboard::kNumFormats) {
-			log((CLOG_DEBUG1 "  no format for target %d", actualTarget));
-			continue;
-		}
-		if (m_added[actualFormat]) {
-			log((CLOG_DEBUG1 "  skipping handled format %d", actualFormat));
 			continue;
 		}
 
 		// add to clipboard and note we've done it
-		m_data[actualFormat]  = targetData;
+		m_data[converter->getFormat()]  = converter->toIClipboard(targetData);
-		m_added[actualFormat] = true;
+		m_added[converter->getFormat()] = true;
-		log((CLOG_DEBUG "  added format %d for target %d", actualFormat, target));
+		log((CLOG_DEBUG "  added format %d for target %d", converter->getFormat(), target));
 	}
 }
 
@@ -676,42 +711,36 @@ CXWindowsClipboard::motifFillCache()
 			continue;
 		}
 
-		// determine the expected clipboard format
-		Atom target = motifFormat->m_type;
-		IClipboard::EFormat expectedFormat = getFormat(target);
-		if (expectedFormat == IClipboard::kNumFormats) {
-			log((CLOG_DEBUG1 "  no format for target %d", target));
-			continue;
-		}
-		log((CLOG_DEBUG1 " source target %d -> %d", target, expectedFormat));
 
-		// skip already handled targets
+		// see if we have a converter for this target
-		if (m_added[expectedFormat]) {
+		Atom target = motifFormat->m_type;
-			log((CLOG_DEBUG1 "  skipping handled format %d", expectedFormat));
+		IXWindowsClipboardConverter* converter = getConverter(target, true);
+		if (converter == NULL) {
 			continue;
 		}
 
 		// get the data (finally)
 		SInt32 length = motifFormat->m_length;
 		sprintf(name, "_MOTIF_CLIP_ITEM_%d", motifFormat->m_data);
-    	Atom atomData = XInternAtom(m_display, name, False);
+    	Atom atomData = XInternAtom(m_display, name, False), atomTarget;
-		data = "";
+		CString targetData;
 		if (!CXWindowsUtil::getWindowProperty(m_display, root,
-									atomData, &data,
+									atomData, &targetData,
-									&target, &format, False)) {
+									&atomTarget, &format, False)) {
+			log((CLOG_DEBUG1 "  no data for target", target));
 			continue;
 		}
-		if (target != atomData) {
+		if (atomTarget != atomData) {
 			continue;
 		}
 
 		// truncate data to length specified in the format
-		data.erase(length);
+		targetData.erase(length);
 
 		// add to clipboard and note we've done it
-		m_data[expectedFormat]  = data;
+		m_data[converter->getFormat()]  = converter->toIClipboard(targetData);
-		m_added[expectedFormat] = true;
+		m_added[converter->getFormat()] = true;
-		log((CLOG_DEBUG "  added format %d for target %d", expectedFormat, motifFormat->m_type));
+		log((CLOG_DEBUG "  added format %d for target %d", converter->getFormat(), target));
 	}
 }
 
@@ -1061,7 +1090,7 @@ CXWindowsClipboard::getTargetsData(CString& data, int* format) const
 {
 	assert(format != NULL);
 
-	// construct response
+	// add standard targets
 	Atom atom;
 	atom = m_atomTargets;
 	data.append(reinterpret_cast<char*>(&atom), sizeof(Atom));
@@ -1069,11 +1098,17 @@ CXWindowsClipboard::getTargetsData(CString& data, int* format) const
 	data.append(reinterpret_cast<char*>(&atom), sizeof(Atom));
 	atom = m_atomTimestamp;
 	data.append(reinterpret_cast<char*>(&atom), sizeof(Atom));
-	if (m_added[kText]) {
+
-		atom = m_atomString;
+	// add targets we can convert to
-		data.append(reinterpret_cast<char*>(&atom), sizeof(Atom));
+	for (ConverterList::const_iterator index = m_converters.begin();
-		atom = m_atomText;
+								index != m_converters.end(); ++index) {
-		data.append(reinterpret_cast<char*>(&atom), sizeof(Atom));
+		IXWindowsClipboardConverter* converter = *index;
+
+		// skip formats we don't have
+		if (!m_added[converter->getFormat()]) {
+			atom = converter->getAtom();
+			data.append(reinterpret_cast<char*>(&atom), sizeof(Atom));
+		}
 	}
 
 	*format = 32;
@@ -1092,21 +1127,6 @@ CXWindowsClipboard::getTimestampData(CString& data, int* format) const
 	return m_atomTimestamp;
 }
 
-Atom
-CXWindowsClipboard::getStringData(CString& data, int* format) const
-{
-	assert(format != NULL);
-
-	if (m_added[kText]) {
-		data    = m_data[kText];
-		*format = 8;
-		return m_atomString;
-	}
-	else {
-		return None;
-	}
-}
-
 
 //
 // CXWindowsClipboard::CICCCMGetClipboard

platform/CXWindowsClipboard.h

@@ -5,12 +5,15 @@
 #include "ClipboardTypes.h"
 #include "stdmap.h"
 #include "stdlist.h"
+#include "stdvector.h"
 #if defined(X_DISPLAY_MISSING)
 #	error X11 is required to build synergy
 #else
 #	include <X11/Xlib.h>
 #endif
 
+class IXWindowsClipboardConverter;
+
 class CXWindowsClipboard : public IClipboard {
 public:
 	CXWindowsClipboard(Display*, Window, ClipboardID);
@@ -51,6 +54,17 @@ public:
 	virtual CString		get(EFormat) const;
 
 private:
+	// remove all converters from our list
+	void				clearConverters();
+
+	// get the converter for a clipboard format.  returns NULL if no
+	// suitable converter.  iff onlyIfNotAdded is true then also
+	// return NULL if a suitable converter was found but we already
+	// have data of the converter's clipboard format.
+	IXWindowsClipboardConverter*
+						getConverter(Atom target,
+							bool onlyIfNotAdded = false) const;
+
 	// convert target atom to clipboard format
 	EFormat				getFormat(Atom target) const;
 
@@ -218,9 +232,10 @@ private:
 	// data conversion methods
 	Atom				getTargetsData(CString&, int* format) const;
 	Atom				getTimestampData(CString&, int* format) const;
-	Atom				getStringData(CString&, int* format) const;
 
 private:
+	typedef std::vector<IXWindowsClipboardConverter*> ConverterList;
+
 	Display*			m_display;
 	Window				m_window;
 	ClipboardID			m_id;
@@ -245,6 +260,9 @@ private:
 	CReplyMap			m_replies;
 	CReplyEventMask		m_eventMasks;
 
+	// clipboard format converters
+	ConverterList		m_converters;
+
 	// atoms we'll need
 	Atom				m_atomTargets;
 	Atom				m_atomMultiple;
@@ -263,4 +281,31 @@ private:
 	Atom				m_atomGDKSelection;
 };
 
+class IXWindowsClipboardConverter : public IInterface {
+public:
+	// accessors
+
+	// return the clipboard format this object converts from/to
+	virtual IClipboard::EFormat
+						getFormat() const = 0;
+
+	// return the atom representing the X selection format that
+	// this object converts from/to
+	virtual Atom		getAtom() const = 0;
+
+	// return the size (in bits) of data elements returned by
+	// toIClipboard().
+	virtual int			getDataSize() const = 0;
+
+	// convert from the IClipboard format to the X selection format.
+	// the input data must be in the IClipboard format returned by
+	// getFormat().  the return data will be in the X selection
+	// format returned by getAtom().
+	virtual CString		fromIClipboard(const CString&) const = 0;
+
+	// convert from the X selection format to the IClipboard format
+	// (i.e., the reverse of fromIClipboard()).
+	virtual CString		toIClipboard(const CString&) const = 0;
+};
+
 #endif

platform/CXWindowsClipboardTextConverter.cpp

@@ -0,0 +1,48 @@
+#include "CXWindowsClipboardTextConverter.h"
+#include "CUnicode.h"
+
+//
+// CXWindowsClipboardTextConverter
+//
+
+CXWindowsClipboardTextConverter::CXWindowsClipboardTextConverter(
+				Display* display, const char* name) :
+	m_atom(XInternAtom(display, name, False))
+{
+	// do nothing
+}
+
+CXWindowsClipboardTextConverter::~CXWindowsClipboardTextConverter()
+{
+	// do nothing
+}
+
+IClipboard::EFormat
+CXWindowsClipboardTextConverter::getFormat() const
+{
+	return IClipboard::kText;
+}
+
+Atom
+CXWindowsClipboardTextConverter::getAtom() const
+{
+	return m_atom;
+}
+
+int
+CXWindowsClipboardTextConverter::getDataSize() const
+{
+	return 8;
+}
+
+CString
+CXWindowsClipboardTextConverter::fromIClipboard(const CString& data) const
+{
+	return CUnicode::UTF8ToText(data);
+}
+
+CString
+CXWindowsClipboardTextConverter::toIClipboard(const CString& data) const
+{
+	return CUnicode::textToUTF8(data);
+}

platform/CXWindowsClipboardTextConverter.h

@@ -0,0 +1,23 @@
+#ifndef CXWINDOWSCLIPBOARDTEXTCONVERTER_H
+#define CXWINDOWSCLIPBOARDTEXTCONVERTER_H
+
+#include "CXWindowsClipboard.h"
+
+class CXWindowsClipboardTextConverter : public IXWindowsClipboardConverter {
+public:
+	CXWindowsClipboardTextConverter(Display* display, const char* name);
+	virtual ~CXWindowsClipboardTextConverter();
+
+	// IXWindowsClipboardConverter overrides
+	virtual IClipboard::EFormat
+						getFormat() const;
+	virtual Atom		getAtom() const;
+	virtual int			getDataSize() const;
+	virtual CString		fromIClipboard(const CString&) const;
+	virtual CString		toIClipboard(const CString&) const;
+
+private:
+	Atom				m_atom;
+};
+
+#endif

platform/CXWindowsClipboardUCS2Converter.cpp

@@ -0,0 +1,48 @@
+#include "CXWindowsClipboardUCS2Converter.h"
+#include "CUnicode.h"
+
+//
+// CXWindowsClipboardUCS2Converter
+//
+
+CXWindowsClipboardUCS2Converter::CXWindowsClipboardUCS2Converter(
+				Display* display, const char* name) :
+	m_atom(XInternAtom(display, name, False))
+{
+	// do nothing
+}
+
+CXWindowsClipboardUCS2Converter::~CXWindowsClipboardUCS2Converter()
+{
+	// do nothing
+}
+
+IClipboard::EFormat
+CXWindowsClipboardUCS2Converter::getFormat() const
+{
+	return IClipboard::kText;
+}
+
+Atom
+CXWindowsClipboardUCS2Converter::getAtom() const
+{
+	return m_atom;
+}
+
+int
+CXWindowsClipboardUCS2Converter::getDataSize() const
+{
+	return 16;
+}
+
+CString
+CXWindowsClipboardUCS2Converter::fromIClipboard(const CString& data) const
+{
+	return CUnicode::UTF8ToUCS2(data);
+}
+
+CString
+CXWindowsClipboardUCS2Converter::toIClipboard(const CString& data) const
+{
+	return CUnicode::UCS2ToUTF8(data);
+}

platform/CXWindowsClipboardUCS2Converter.h

@@ -0,0 +1,23 @@
+#ifndef CXWINDOWSCLIPBOARDUCS2CONVERTER_H
+#define CXWINDOWSCLIPBOARDUCS2CONVERTER_H
+
+#include "CXWindowsClipboard.h"
+
+class CXWindowsClipboardUCS2Converter : public IXWindowsClipboardConverter {
+public:
+	CXWindowsClipboardUCS2Converter(Display* display, const char* name);
+	virtual ~CXWindowsClipboardUCS2Converter();
+
+	// IXWindowsClipboardConverter overrides
+	virtual IClipboard::EFormat
+						getFormat() const;
+	virtual Atom		getAtom() const;
+	virtual int			getDataSize() const;
+	virtual CString		fromIClipboard(const CString&) const;
+	virtual CString		toIClipboard(const CString&) const;
+
+private:
+	Atom				m_atom;
+};
+
+#endif

platform/CXWindowsClipboardUTF8Converter.cpp

@@ -0,0 +1,47 @@
+#include "CXWindowsClipboardUTF8Converter.h"
+
+//
+// CXWindowsClipboardUTF8Converter
+//
+
+CXWindowsClipboardUTF8Converter::CXWindowsClipboardUTF8Converter(
+				Display* display, const char* name) :
+	m_atom(XInternAtom(display, name, False))
+{
+	// do nothing
+}
+
+CXWindowsClipboardUTF8Converter::~CXWindowsClipboardUTF8Converter()
+{
+	// do nothing
+}
+
+IClipboard::EFormat
+CXWindowsClipboardUTF8Converter::getFormat() const
+{
+	return IClipboard::kText;
+}
+
+Atom
+CXWindowsClipboardUTF8Converter::getAtom() const
+{
+	return m_atom;
+}
+
+int
+CXWindowsClipboardUTF8Converter::getDataSize() const
+{
+	return 8;
+}
+
+CString
+CXWindowsClipboardUTF8Converter::fromIClipboard(const CString& data) const
+{
+	return data;
+}
+
+CString
+CXWindowsClipboardUTF8Converter::toIClipboard(const CString& data) const
+{
+	return data;
+}

platform/CXWindowsClipboardUTF8Converter.h

@@ -0,0 +1,23 @@
+#ifndef CXWINDOWSCLIPBOARDUTF8CONVERTER_H
+#define CXWINDOWSCLIPBOARDUTF8CONVERTER_H
+
+#include "CXWindowsClipboard.h"
+
+class CXWindowsClipboardUTF8Converter : public IXWindowsClipboardConverter {
+public:
+	CXWindowsClipboardUTF8Converter(Display* display, const char* name);
+	virtual ~CXWindowsClipboardUTF8Converter();
+
+	// IXWindowsClipboardConverter overrides
+	virtual IClipboard::EFormat
+						getFormat() const;
+	virtual Atom		getAtom() const;
+	virtual int			getDataSize() const;
+	virtual CString		fromIClipboard(const CString&) const;
+	virtual CString		toIClipboard(const CString&) const;
+
+private:
+	Atom				m_atom;
+};
+
+#endif

platform/Makefile.am

@@ -4,22 +4,28 @@ DEPTH = ..
 
 # FIXME -- add CUnixPlatform.cpp as an unbuilt source
 noinst_LIBRARIES = libplatform.a
-libplatform_a_SOURCES = 		\
+libplatform_a_SOURCES = 					\
-	CPlatform.cpp				\
+	CPlatform.cpp							\
-	CXWindowsClipboard.cpp		\
+	CXWindowsClipboard.cpp					\
-	CXWindowsScreen.cpp			\
+	CXWindowsClipboardTextConverter.cpp		\
-	CXWindowsScreenSaver.cpp	\
+	CXWindowsClipboardUCS2Converter.cpp		\
-	CXWindowsUtil.cpp			\
+	CXWindowsClipboardUTF8Converter.cpp		\
-	CPlatform.h					\
+	CXWindowsScreen.cpp						\
-	CUnixPlatform.h				\
+	CXWindowsScreenSaver.cpp				\
-	CXWindowsClipboard.h		\
+	CXWindowsUtil.cpp						\
-	CXWindowsScreen.h			\
+	CPlatform.h								\
-	CXWindowsScreenSaver.h		\
+	CUnixPlatform.h							\
-	CXWindowsUtil.h				\
+	CXWindowsClipboard.h					\
-	IPlatform.h					\
+	CXWindowsClipboardTextConverter.h		\
+	CXWindowsClipboardUCS2Converter.h		\
+	CXWindowsClipboardUTF8Converter.h		\
+	CXWindowsScreen.h						\
+	CXWindowsScreenSaver.h					\
+	CXWindowsUtil.h							\
+	IPlatform.h								\
 	$(NULL)
-INCLUDES =					\
+INCLUDES =									\
-	-I$(DEPTH)/base 		\
+	-I$(DEPTH)/base 						\
-	-I$(DEPTH)/mt	 		\
+	-I$(DEPTH)/mt	 						\
-	-I$(DEPTH)/synergy 		\
+	-I$(DEPTH)/synergy 						\
 	$(NULL)

synergy/IClipboard.h

@@ -13,7 +13,13 @@ public:
 	typedef UInt32 Time;
 
 	// known clipboard formats.  kNumFormats must be last and
-	// formats must be sequential starting from zero.
+	// formats must be sequential starting from zero.  clipboard
+	// data set via add() and retrieved via get() is in one of
+	// these formats.  platform dependent clipboard subclasses
+	// can and should present any suitable formats derivable
+	// from these formats (e.g. UCS-16 encoded unicode).
+	//
+	// kText:  UTF-8 encoded unicode (ISO-10646), newline is LF
 	enum EFormat { kText, kNumFormats };
 
 	// manipulators