Merge 8.6

3 files changed, 70 insertions, 35 deletions

diff --git a/unix/tkUnixFont.c b/unix/tkUnixFont.c
index 908deeb..3c1ce65 100644
--- a/unix/tkUnixFont.c
+++ b/unix/tkUnixFont.c
@@ -35,10 +35,10 @@ static const char encodingList[][10] = {
  * family": the foundry, face name, and charset.
  */
 
-#define FONTMAP_SHIFT		12
+#define FONTMAP_SHIFT		10
 
-#define FONTMAP_PAGES		(1 << (21 - FONTMAP_SHIFT))
 #define FONTMAP_BITSPERPAGE	(1 << FONTMAP_SHIFT)
+#define FONTMAP_PAGES		(0x30000 / FONTMAP_BITSPERPAGE)
 
 typedef struct FontFamily {
     struct FontFamily *nextPtr;	/* Next in list of all known font families. */
@@ -391,7 +391,7 @@ ControlUtfProc(
     const char *srcStart, *srcEnd;
     char *dstStart, *dstEnd;
     int ch, result;
-    static const char hexChars[] = "0123456789abcdef";
+    static const char hexChars[] = "0123456789ABCDEF";
     static const char mapChars[] = {
 	0, 0, 0, 0, 0, 0, 0,
 	'a', 'b', 't', 'n', 'v', 'f', 'r'
@@ -412,15 +412,15 @@ ControlUtfProc(
 	}
 	src += TkUtfToUniChar(src, &ch);
 	dst[0] = '\\';
-	if (((size_t) 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) {
+	} else if ((size_t)ch < 256) {
 	    dst[1] = 'x';
 	    dst[2] = hexChars[(ch >> 4) & 0xf];
 	    dst[3] = hexChars[ch & 0xf];
 	    dst += 4;
-	} else if (ch < 0x10000) {
+	} else if ((size_t)ch < 0x10000) {
 	    dst[1] = 'u';
 	    dst[2] = hexChars[(ch >> 12) & 0xf];
 	    dst[3] = hexChars[(ch >> 8) & 0xf];
@@ -430,10 +430,10 @@ ControlUtfProc(
 	} else {
 	    /* TODO we can do better here */
 	    dst[1] = 'u';
-	    dst[2] = 'f';
-	    dst[3] = 'f';
-	    dst[4] = 'f';
-	    dst[5] = 'd';
+	    dst[2] = 'F';
+	    dst[3] = 'F';
+	    dst[4] = 'F';
+	    dst[5] = 'D';
 	    dst += 6;
 	}
     }
@@ -449,7 +449,6 @@ ControlUtfProc(
  * Ucs2beToUtfProc --
  *
  *	Convert from UCS-2BE (big-endian 16-bit Unicode) to UTF-8.
- *	This is only defined on LE machines.
  *
  * Results:
  *	Returns TCL_OK if conversion was successful.
@@ -498,6 +497,11 @@ Ucs2beToUtfProc(
 	result = TCL_CONVERT_MULTIBYTE;
 	srcLen--;
     }
+    /* If last code point is a high surrogate, we cannot handle that yet */
+    if ((srcLen >= 2) && ((src[srcLen - 2] & 0xFC) == 0xD8)) {
+	result = TCL_CONVERT_MULTIBYTE;
+	srcLen -= 2;
+    }
 
     srcStart = src;
     srcEnd = src + srcLen;
@@ -572,7 +576,11 @@ UtfToUcs2beProc(
 {
     const char *srcStart, *srcEnd, *srcClose, *dstStart, *dstEnd;
     int result, numChars;
-    Tcl_UniChar ch;
+    Tcl_UniChar *chPtr = (Tcl_UniChar *)statePtr;
+
+    if (flags & TCL_ENCODING_START) {
+	*statePtr = 0;
+    }
 
     srcStart = src;
     srcEnd = src + srcLen;
@@ -591,15 +599,14 @@ UtfToUcs2beProc(
 	     * If there is more string to follow, this will ensure that the
 	     * last UTF-8 character in the source buffer hasn't been cut off.
 	     */
-
 	    result = TCL_CONVERT_MULTIBYTE;
 	    break;
 	}
 	if (dst > dstEnd) {
 	    result = TCL_CONVERT_NOSPACE;
 	    break;
-        }
-	src += Tcl_UtfToUniChar(src, &ch);
+	}
+	src += Tcl_UtfToUniChar(src, chPtr);
 
 	/*
 	 * Ensure big-endianness (store big bits first).
@@ -607,8 +614,9 @@ UtfToUcs2beProc(
 	 * sure to work in char* for Tcl_UtfToUniChar alignment. [Bug 1122671]
 	 */
 
-	*dst++ = (ch >> 8);
-	*dst++ = (ch & 0xFF);
+
+	*dst++ = (char)(*chPtr >> 8);
+	*dst++ = (char)*chPtr;
     }
     *srcReadPtr = src - srcStart;
     *dstWrotePtr = dst - dstStart;
@@ -1967,11 +1975,11 @@ FindSubFontForChar(
     SubFont *subFontPtr;
     Tcl_DString ds;
 
-    if (FontMapLookup(&fontPtr->subFontArray[0], ch)) {
-	return &fontPtr->subFontArray[0];
+    if (ch < 0 || ch > 0x30000) {
+	ch = 0xfffd;
     }
 
-    for (i = 1; i < fontPtr->numSubFonts; i++) {
+    for (i = 0; i < fontPtr->numSubFonts; i++) {
 	if (FontMapLookup(&fontPtr->subFontArray[i], ch)) {
 	    return &fontPtr->subFontArray[i];
 	}
@@ -2121,6 +2129,9 @@ FontMapLookup(
 {
     int row, bitOffset;
 
+    if (ch < 0 ||  ch >= 0x30000) {
+	return 0;
+    }
     row = ch >> FONTMAP_SHIFT;
     if (subFontPtr->fontMap[row] == NULL) {
 	FontMapLoadPage(subFontPtr, row);
@@ -2161,12 +2172,14 @@ FontMapInsert(
 {
     int row, bitOffset;
 
-    row = ch >> FONTMAP_SHIFT;
-    if (subFontPtr->fontMap[row] == NULL) {
-	FontMapLoadPage(subFontPtr, row);
+    if (ch >= 0 &&  ch < 0x30000) {
+	row = ch >> FONTMAP_SHIFT;
+	if (subFontPtr->fontMap[row] == NULL) {
+	    FontMapLoadPage(subFontPtr, row);
+	}
+	bitOffset = ch & (FONTMAP_BITSPERPAGE - 1);
+	subFontPtr->fontMap[row][bitOffset >> 3] |= 1 << (bitOffset & 7);
     }
-    bitOffset = ch & (FONTMAP_BITSPERPAGE - 1);
-    subFontPtr->fontMap[row][bitOffset >> 3] |= 1 << (bitOffset & 7);
 }
 
 /*
diff --git a/unix/tkUnixSelect.c b/unix/tkUnixSelect.c
index b6e980e..11161b7 100644
--- a/unix/tkUnixSelect.c
+++ b/unix/tkUnixSelect.c
@@ -21,7 +21,7 @@ typedef struct ConvertInfo {
 				 * offset of the next chunk of data to
 				 * transfer. */
     Tcl_EncodingState state;	/* The encoding state needed across chunks. */
-    char buffer[4];		/* A buffer to hold part of a UTF character
+    char buffer[4];	/* A buffer to hold part of a UTF character
 				 * that is split across chunks.*/
 } ConvertInfo;
 
@@ -446,7 +446,7 @@ TkSelPropProc(
 		 * Preserve any left-over bytes.
 		 */
 
-		if (srcLen > 4) {
+		if (srcLen > 3) {
 		    Tcl_Panic("selection conversion left too many bytes unconverted");
 		}
 		memcpy(incrPtr->converts[i].buffer, src, srcLen + 1);
diff --git a/win/tkWinFont.c b/win/tkWinFont.c
index 948ca10..35a9941 100644
--- a/win/tkWinFont.c
+++ b/win/tkWinFont.c
@@ -26,10 +26,10 @@
  * Under Windows, a "font family" is uniquely identified by its face name.
  */
 
-#define FONTMAP_SHIFT	    12
+#define FONTMAP_SHIFT	    10
 
-#define FONTMAP_PAGES	    	(1 << (21 - FONTMAP_SHIFT))
 #define FONTMAP_BITSPERPAGE	(1 << FONTMAP_SHIFT)
+#define FONTMAP_PAGES		(0x30000 / FONTMAP_BITSPERPAGE)
 
 typedef struct FontFamily {
     struct FontFamily *nextPtr;	/* Next in list of all known font families. */
@@ -1943,7 +1943,7 @@ FindSubFontForChar(
     SubFont *subFontPtr;
     Tcl_DString ds;
 
-    if (ch < BASE_CHARS) {
+    if ((ch < BASE_CHARS) || (ch >= 0x30000)) {
 	return &fontPtr->subFontArray[0];
     }
 
@@ -2115,6 +2115,10 @@ FontMapLookup(
 {
     int row, bitOffset;
 
+    if (ch < 0 || ch >= 0x30000) {
+	return 0;
+    }
+
     row = ch >> FONTMAP_SHIFT;
     if (subFontPtr->fontMap[row] == NULL) {
 	FontMapLoadPage(subFontPtr, row);
@@ -2155,12 +2159,14 @@ FontMapInsert(
 {
     int row, bitOffset;
 
-    row = ch >> FONTMAP_SHIFT;
-    if (subFontPtr->fontMap[row] == NULL) {
-	FontMapLoadPage(subFontPtr, row);
+    if (ch >= 0 && ch < 0x30000) {
+	row = ch >> FONTMAP_SHIFT;
+	if (subFontPtr->fontMap[row] == NULL) {
+	    FontMapLoadPage(subFontPtr, row);
+	}
+	bitOffset = ch & (FONTMAP_BITSPERPAGE - 1);
+	subFontPtr->fontMap[row][bitOffset >> 3] |= 1 << (bitOffset & 7);
     }
-    bitOffset = ch & (FONTMAP_BITSPERPAGE - 1);
-    subFontPtr->fontMap[row][bitOffset >> 3] |= 1 << (bitOffset & 7);
 }
 
 /*
@@ -2528,6 +2534,22 @@ FamilyExists(
     int result;
     Tcl_DString faceString;
 
+    /*
+     * Just immediately rule out the following fonts, because they look so
+     * ugly on windows. The caller's fallback mechanism will cause the
+     * corresponding appropriate TrueType fonts to be selected.
+     */
+
+    if (strcasecmp(faceName, "Courier") == 0) {
+	return 0;
+    }
+    if (strcasecmp(faceName, "Times") == 0) {
+	return 0;
+    }
+    if (strcasecmp(faceName, "Helvetica") == 0) {
+	return 0;
+    }
+
     Tcl_UtfToExternalDString(systemEncoding, faceName, -1, &faceString);
 
     /*