More simplifications

13 files changed, 89 insertions, 75 deletions

diff --git a/generic/tkEntry.c b/generic/tkEntry.c
index a66cf18..34f11d2 100644
--- a/generic/tkEntry.c
+++ b/generic/tkEntry.c
@@ -1926,8 +1926,8 @@ EntryComputeGeometry(
      */
 
     if (entryPtr->showChar != NULL) {
-	Tcl_UniChar ch;
-	char buf[4];
+	int ch;
+	char buf[6];
 	int size;
 
 	/*
@@ -1937,8 +1937,8 @@ EntryComputeGeometry(
 	 * resulting string.
 	 */
 
-	Tcl_UtfToUniChar(entryPtr->showChar, &ch);
-	size = Tcl_UniCharToUtf(ch, buf);
+	TkUtfToUniChar(entryPtr->showChar, &ch);
+	size = TkUniCharToUtf(ch, buf);
 
 	entryPtr->numDisplayBytes = entryPtr->numChars * size;
 	p = ckalloc(entryPtr->numDisplayBytes + 1);
diff --git a/generic/tkFont.c b/generic/tkFont.c
index 4a45691..1947666 100644
--- a/generic/tkFont.c
+++ b/generic/tkFont.c
@@ -563,7 +563,7 @@ Tk_FontObjCmd(
 
 	if (charPtr != NULL) {
 	    const char *string = Tcl_GetString(charPtr);
-	    int len = TkUtfToUniChar2(string, &uniChar);
+	    int len = TkUtfToUniChar(string, &uniChar);
 
 	    if (len != charPtr->length) {
 		resultPtr = Tcl_NewStringObj(
@@ -1714,7 +1714,7 @@ Tk_PostscriptFontName(
 		src++;
 		upper = 1;
 	    }
-	    src += TkUtfToUniChar2(src, &ch);
+	    src += TkUtfToUniChar(src, &ch);
 	    if (upper) {
 		ch = (Tcl_UniChar) Tcl_UniCharToUpper(ch);
 		upper = 0;
@@ -3274,7 +3274,7 @@ Tk_TextLayoutToPostscript(
 	     * international postscript fonts.
 	     */
 
-	    p += TkUtfToUniChar2(p, &ch);
+	    p += TkUtfToUniChar(p, &ch);
 	    if ((ch == '(') || (ch == ')') || (ch == '\\') || (ch < 0x20)) {
 		/*
 		 * Tricky point: the "03" is necessary in the sprintf below,
diff --git a/generic/tkInt.h b/generic/tkInt.h
index 6d86e08..1615a81 100644
--- a/generic/tkInt.h
+++ b/generic/tkInt.h
@@ -1233,9 +1233,11 @@ MODULE_SCOPE void	TkUnixSetXftClipRegion(TkRegion clipRegion);
 #endif
 
 #if TCL_UTF_MAX > 4
-#   define TkUtfToUniChar2 Tcl_UtfToUniChar
+#   define TkUtfToUniChar Tcl_UtfToUniChar
+#   define TkUniCharToUtf Tcl_UniCharToUtf
 #else
-    MODULE_SCOPE int TkUtfToUniChar2(const char *src, int *chPtr);
+    MODULE_SCOPE int TkUtfToUniChar(const char *, int *);
+    MODULE_SCOPE int TkUniCharToUtf(int, char *);
 #endif
 
 /*
diff --git a/generic/tkText.c b/generic/tkText.c
index dacadbe..8ae17a5 100644
--- a/generic/tkText.c
+++ b/generic/tkText.c
@@ -4615,7 +4615,7 @@ TkTextGetTabs(
 	 * There may be a more efficient way of getting this.
 	 */
 
-	TkUtfToUniChar2(Tcl_GetString(objv[i+1]), &ch);
+	TkUtfToUniChar(Tcl_GetString(objv[i+1]), &ch);
 	if (!Tcl_UniCharIsAlpha(ch)) {
 	    continue;
 	}
@@ -5966,7 +5966,7 @@ SearchCore(
 
 	    CLANG_ASSERT(pattern);
 	    do {
-		Tcl_UniChar ch;
+		int ch;
 		const char *p;
 		int lastFullLine = lastOffset;
 
@@ -6196,7 +6196,7 @@ SearchCore(
 			}
 		    } else {
 			firstOffset = p - startOfLine +
-				Tcl_UtfToUniChar(startOfLine+matchOffset,&ch);
+				TkUtfToUniChar(startOfLine+matchOffset,&ch);
 		    }
 		}
 	    } while (searchSpecPtr->all);
diff --git a/generic/tkTextDisp.c b/generic/tkTextDisp.c
index 026023e..eb917cf 100644
--- a/generic/tkTextDisp.c
+++ b/generic/tkTextDisp.c
@@ -7582,7 +7582,7 @@ TkTextCharLayoutProc(
     if (bytesThatFit < maxBytes) {
 	if ((bytesThatFit == 0) && noCharsYet) {
 	    int ch;
-	    int chLen = TkUtfToUniChar2(p, &ch);
+	    int chLen = TkUtfToUniChar(p, &ch);
 
 #if TK_LAYOUT_WITH_BASE_CHUNKS
 	    bytesThatFit = CharChunkMeasureChars(chunkPtr, line,
diff --git a/generic/tkTextIndex.c b/generic/tkTextIndex.c
index b794cdb..7aebc84 100644
--- a/generic/tkTextIndex.c
+++ b/generic/tkTextIndex.c
@@ -2300,7 +2300,7 @@ StartEnd(
 	    if (segPtr->typePtr == &tkTextCharType) {
 		int ch;
 
-		chSize = TkUtfToUniChar2(segPtr->body.chars + offset, &ch);
+		chSize = TkUtfToUniChar(segPtr->body.chars + offset, &ch);
 		if (!Tcl_UniCharIsWordChar(ch)) {
 		    break;
 		}
@@ -2345,7 +2345,7 @@ StartEnd(
 	    if (segPtr->typePtr == &tkTextCharType) {
 
 		int ch;
-		TkUtfToUniChar2(segPtr->body.chars + offset, &ch);
+		TkUtfToUniChar(segPtr->body.chars + offset, &ch);
 		if (!Tcl_UniCharIsWordChar(ch)) {
 		    break;
 		}
diff --git a/generic/tkUtil.c b/generic/tkUtil.c
index a266cb3..19b343e 100644
--- a/generic/tkUtil.c
+++ b/generic/tkUtil.c
@@ -1197,7 +1197,7 @@ TkSendVirtualEvent(
 /*
  *---------------------------------------------------------------------------
  *
- * TkUtfToUniChar2 --
+ * TkUtfToUniChar --
  *
  *	Almost the same as Tcl_UtfToUniChar but using int instead of Tcl_UniChar.
  *	This function is capable of collapsing a upper/lower pair to a single
@@ -1214,7 +1214,7 @@ TkSendVirtualEvent(
  */
 
 int
-TkUtfToUniChar2(
+TkUtfToUniChar(
     const char *src,	/* The UTF-8 string. */
     int *chPtr)		/* Filled with the Tcl_UniChar represented by
 			 * the UTF-8 string. */
@@ -1238,6 +1238,39 @@ TkUtfToUniChar2(
     }
     return len;
 }
+
+/*
+ *---------------------------------------------------------------------------
+ *
+ * TkUniCharToUtf --
+ *
+ *	Almost the same as Tcl_UniCharToUtf but producing surrogates if
+ *	TCL_UTF_MAX==3.
+ *
+ * Results:
+ *	*buf is filled with the UTF-8 string, and the return value is the
+ *	number of bytes produced.
+ *
+ * Side effects:
+ *	None.
+ *
+ *---------------------------------------------------------------------------
+ */
+
+int TkUniCharToUtf(int ch, char *buf)
+{
+    int size = Tcl_UniCharToUtf(ch, buf);
+    if ((ch > 0xffff) && (size < 4)) {
+	/* Hey, this is wrong, we must be running TCL_UTF_MAX==3
+	 * The best thing we can do is spit out 2 surrogates */
+	ch -= 0x10000;
+	size = Tcl_UniCharToUtf(((ch >> 10) | 0xd800), buf);
+	size += Tcl_UniCharToUtf(((ch & 0x3ff) | 0xdc00), buf+size);
+    }
+    return size;
+}
+
+
 #endif
 /*
  * Local Variables:
diff --git a/generic/ttk/ttkEntry.c b/generic/ttk/ttkEntry.c
index d80e1fd..a25574a 100644
--- a/generic/ttk/ttkEntry.c
+++ b/generic/ttk/ttkEntry.c
@@ -282,11 +282,11 @@ static char *EntryDisplayString(const char *showChar, int numChars)
 {
     char *displayString, *p;
     int size;
-    Tcl_UniChar ch;
-    char buf[4];
+    int ch;
+    char buf[6];
 
-    Tcl_UtfToUniChar(showChar, &ch);
-    size = Tcl_UniCharToUtf(ch, buf);
+    TkUtfToUniChar(showChar, &ch);
+    size = TkUniCharToUtf(ch, buf);
     p = displayString = ckalloc(numChars * size + 1);
 
     while (numChars--) {
@@ -406,7 +406,7 @@ ExpandPercents(
     int number, length;
     const char *string;
     int stringLength;
-    Tcl_UniChar ch;
+    int ch;
     char numStorage[2*TCL_INTEGER_SPACE];
 
     while (*template) {
@@ -430,7 +430,7 @@ ExpandPercents(
 	 */
 	++template; /* skip over % */
 	if (*template != '\0') {
-	    template += Tcl_UtfToUniChar(template, &ch);
+	    template += TkUtfToUniChar(template, &ch);
 	} else {
 	    ch = '%';
 	}
@@ -480,7 +480,7 @@ ExpandPercents(
 		string = Tk_PathName(entryPtr->core.tkwin);
 		break;
 	    default:
-		length = Tcl_UniCharToUtf(ch, numStorage);
+		length = TkUniCharToUtf(ch, numStorage);
 		numStorage[length] = '\0';
 		string = numStorage;
 		break;
diff --git a/unix/tkUnixFont.c b/unix/tkUnixFont.c
index 681d1d1..ce4eca9 100644
--- a/unix/tkUnixFont.c
+++ b/unix/tkUnixFont.c
@@ -405,7 +405,7 @@ ControlUtfProc(
 {
     const char *srcStart, *srcEnd;
     char *dstStart, *dstEnd;
-    Tcl_UniChar ch;
+    int ch;
     int result;
     static char hexChars[] = "0123456789abcdef";
     static char mapChars[] = {
@@ -426,9 +426,9 @@ ControlUtfProc(
 	    result = TCL_CONVERT_NOSPACE;
 	    break;
 	}
-	src += Tcl_UtfToUniChar(src, &ch);
+	src += TkUtfToUniChar(src, &ch);
 	dst[0] = '\\';
-	if ((ch < sizeof(mapChars)) && (mapChars[ch] != 0)) {
+	if (((size_t) ch < sizeof(mapChars)) && (mapChars[ch] != 0)) {
 	    dst[1] = mapChars[ch];
 	    dst += 2;
 	} else if (ch < 256) {
@@ -436,13 +436,21 @@ ControlUtfProc(
 	    dst[2] = hexChars[(ch >> 4) & 0xf];
 	    dst[3] = hexChars[ch & 0xf];
 	    dst += 4;
-	} else {
+	} else if (ch < 0x10000) {
 	    dst[1] = 'u';
 	    dst[2] = hexChars[(ch >> 12) & 0xf];
 	    dst[3] = hexChars[(ch >> 8) & 0xf];
 	    dst[4] = hexChars[(ch >> 4) & 0xf];
 	    dst[5] = hexChars[ch & 0xf];
 	    dst += 6;
+	} else {
+	    /* TODO we can do better here */
+	    dst[1] = 'u';
+	    dst[2] = 'f';
+	    dst[3] = 'f';
+	    dst[4] = 'f';
+	    dst[5] = 'd';
+	    dst += 6;
 	}
     }
     *srcReadPtr = src - srcStart;
@@ -1027,7 +1035,7 @@ Tk_MeasureChars(
 	curByte = 0;
     } else if (maxLength < 0) {
 	const char *p, *end, *next;
-	Tcl_UniChar ch;
+	int ch;
 	SubFont *thisSubFontPtr;
 	FontFamily *familyPtr;
 	Tcl_DString runString;
@@ -1043,7 +1051,7 @@ Tk_MeasureChars(
 	curX = 0;
 	end = source + numBytes;
 	for (p = source; p < end; ) {
-	    next = p + Tcl_UtfToUniChar(p, &ch);
+	    next = p + TkUtfToUniChar(p, &ch);
 	    thisSubFontPtr = FindSubFontForChar(fontPtr, ch, &lastSubFontPtr);
 	    if (thisSubFontPtr != lastSubFontPtr) {
 		familyPtr = lastSubFontPtr->familyPtr;
diff --git a/unix/tkUnixRFont.c b/unix/tkUnixRFont.c
index 0ea1ec6..cf4127d 100644
--- a/unix/tkUnixRFont.c
+++ b/unix/tkUnixRFont.c
@@ -670,7 +670,7 @@ Tk_MeasureChars(
     while (numBytes > 0) {
 	int unichar;
 
-	clen = TkUtfToUniChar2(source, &unichar);
+	clen = TkUtfToUniChar(source, &unichar);
 	c = (FcChar32) unichar;
 
 	if (clen <= 0) {
diff --git a/win/tkWinFont.c b/win/tkWinFont.c
index 30638ca..7c6c0ba 100644
--- a/win/tkWinFont.c
+++ b/win/tkWinFont.c
@@ -858,7 +858,7 @@ Tk_MeasureChars(
     start = source;
     end = start + numBytes;
     for (p = start; p < end; ) {
-	next = p + TkUtfToUniChar2(p, &ch);
+	next = p + TkUtfToUniChar(p, &ch);
 	thisSubFontPtr = FindSubFontForChar(fontPtr, ch, &lastSubFontPtr);
 	if (thisSubFontPtr != lastSubFontPtr) {
 	    familyPtr = lastSubFontPtr->familyPtr;
@@ -920,7 +920,7 @@ Tk_MeasureChars(
 	familyPtr = lastSubFontPtr->familyPtr;
 	Tcl_DStringInit(&runString);
 	for (p = start; p < end; ) {
-	    next = p + TkUtfToUniChar2(p, &ch);
+	    next = p + TkUtfToUniChar(p, &ch);
 	    Tcl_UtfToExternal(NULL, familyPtr->encoding, p,
 		    (int) (next - p), 0, NULL, buf, sizeof(buf), NULL,
 		    &dstWrote, NULL);
@@ -975,7 +975,7 @@ Tk_MeasureChars(
 	p = source;
 	ch = ' ';
 	while (p < end) {
-	    next = p + TkUtfToUniChar2(p, &ch2);
+	    next = p + TkUtfToUniChar(p, &ch2);
 	    if ((ch != ' ') && (ch2 == ' ')) {
 		lastWordBreak = p;
 	    }
@@ -1457,7 +1457,7 @@ MultiFontTextOut(
 
     end = source + numBytes;
     for (p = source; p < end; ) {
-	next = p + TkUtfToUniChar2(p, &ch);
+	next = p + TkUtfToUniChar(p, &ch);
 	thisSubFontPtr = FindSubFontForChar(fontPtr, ch, &lastSubFontPtr);
 	if (thisSubFontPtr != lastSubFontPtr) {
 	    if (p > source) {
diff --git a/win/tkWinKey.c b/win/tkWinKey.c
index 2698c4d..7fee101 100644
--- a/win/tkWinKey.c
+++ b/win/tkWinKey.c
@@ -88,6 +88,8 @@ TkpGetString(
 				 * result. */
 {
     XKeyEvent *keyEv = &eventPtr->xkey;
+	int len;
+	char buf[6];
 
     Tcl_DStringInit(dsPtr);
     if (keyEv->send_event == -1) {
@@ -95,41 +97,14 @@ TkpGetString(
 	    Tcl_ExternalToUtfDString(TkWinGetKeyInputEncoding(),
 		    keyEv->trans_chars, keyEv->nbytes, dsPtr);
 	}
-    } else if (keyEv->send_event == -2) {
-	/*
-	 * Special case for win2000 multi-lingal IME input. xkey.trans_chars[]
-	 * already contains a UNICODE char.
-	 */
-
-	int unichar;
-	char buf[XMaxTransChars];
-	int len;
-
-	unichar = keyEv->trans_chars[1] & 0xff;
-	unichar <<= 8;
-	unichar |= keyEv->trans_chars[0] & 0xff;
-
-	len = Tcl_UniCharToUtf((Tcl_UniChar) unichar, buf);
-
-	Tcl_DStringAppend(dsPtr, buf, len);
     } else if (keyEv->send_event == -3) {
 
-	char buf[XMaxTransChars];
-	int len;
-
 	/*
-	 * Special case for WM_UNICHAR.
+	 * Special case for WM_UNICHAR and win2000 multi-lingal IME input
 	 */
 
-	len = Tcl_UniCharToUtf(keyEv->keycode, buf);
-	if ((keyEv->keycode <= 0xffff) || (len == XMaxTransChars)) {
-	    Tcl_DStringAppend(dsPtr, buf, len);
-	} else {
-	    Tcl_UniCharToUtf(((keyEv->keycode - 0x10000) >> 10) | 0xd800, buf);
-	    Tcl_DStringAppend(dsPtr, buf, 3);
-	    Tcl_UniCharToUtf(((keyEv->keycode - 0x10000) & 0x3ff) | 0xdc00, buf);
-	    Tcl_DStringAppend(dsPtr, buf, 3);
-	}
+	len = TkUniCharToUtf(keyEv->keycode, buf);
+	Tcl_DStringAppend(dsPtr, buf, len);
     } else {
 	/*
 	 * This is an event generated from generic code. It has no nchars or
@@ -140,9 +115,6 @@ TkpGetString(
 
 	if (((keysym != NoSymbol) && (keysym > 0) && (keysym < 256))
 		|| (keysym == XK_Return) || (keysym == XK_Tab)) {
-	    char buf[XMaxTransChars];
-	    int len;
-
 	    len = Tcl_UniCharToUtf((Tcl_UniChar) (keysym & 255), buf);
 	    Tcl_DStringAppend(dsPtr, buf, len);
 	}
diff --git a/win/tkWinX.c b/win/tkWinX.c
index 6be54e2..ce73aac 100644
--- a/win/tkWinX.c
+++ b/win/tkWinX.c
@@ -1580,8 +1580,8 @@ HandleIMEComposition(
 	/*
 	 * Set up the fields pertinent to key event.
 	 *
-	 * We set send_event to the special value of -2, so that TkpGetString
-	 * in tkWinKey.c knows that trans_chars[] already contains a UNICODE
+	 * We set send_event to the special value of -3, so that TkpGetString
+	 * in tkWinKey.c knows that keycode already contains a UNICODE
 	 * char and there's no need to do encoding conversion.
 	 *
 	 * Note that the event *must* be zeroed out first; Tk plays cunning
@@ -1592,7 +1592,7 @@ HandleIMEComposition(
 
 	memset(&event, 0, sizeof(XEvent));
 	event.xkey.serial = winPtr->display->request++;
-	event.xkey.send_event = -2;
+	event.xkey.send_event = -3;
 	event.xkey.display = winPtr->display;
 	event.xkey.window = winPtr->window;
 	event.xkey.root = RootWindow(winPtr->display, winPtr->screenNum);
@@ -1600,8 +1600,7 @@ HandleIMEComposition(
 	event.xkey.state = TkWinGetModifierState();
 	event.xkey.time = TkpGetMS();
 	event.xkey.same_screen = True;
-	event.xkey.keycode = 0;
-	event.xkey.nbytes = 2;
+	event.xkey.nbytes = 0;
 
 	for (i=0; i<n; ) {
 	    /*
@@ -1609,8 +1608,8 @@ HandleIMEComposition(
 	     * UNICODE character in the composition.
 	     */
 
-	    event.xkey.trans_chars[0] = (char) buff[i++];
-	    event.xkey.trans_chars[1] = (char) buff[i++];
+	    event.xkey.keycode = ((unsigned char) buff[i++]) << 8;
+	    event.xkey.keycode += (unsigned char) buff[i++];
 
 	    event.type = KeyPress;
 	    Tk_QueueWindowEvent(&event, TCL_QUEUE_TAIL);