Add (and use) new internal function TkUtfAtIndex(), which does the same as Tcl_UtfAtIndex() only it protects against ending in the middle of a 4-byte UTF-8 sequence. This should fix another part of [a179564826] when handling copy-pasted Emoji in Text/Entry (and other) widgets.

3 files changed, 9 insertions, 27 deletions

diff --git a/macosx/tkMacOSXBitmap.c b/macosx/tkMacOSXBitmap.c
index 615192b..2b08235 100644
--- a/macosx/tkMacOSXBitmap.c
+++ b/macosx/tkMacOSXBitmap.c
@@ -317,7 +317,6 @@ TkpGetNativeAppBitmap(
 	OSType iconType;
 	if (OSTypeFromString(name, &iconType) == TCL_OK) {
 	    NSString *iconUTI = OSTYPE_TO_UTI(iconType);
-	    printf("Found image for UTI %s\n", iconUTI.UTF8String);
 	    NSImage *iconImage = [[NSWorkspace sharedWorkspace]
 				     iconForFileType: iconUTI];
 	    pixmap = PixmapFromImage(display, iconImage, NSSizeToCGSize(size));
diff --git a/macosx/tkMacOSXFont.c b/macosx/tkMacOSXFont.c
index 7fc0113..d704184 100644
--- a/macosx/tkMacOSXFont.c
+++ b/macosx/tkMacOSXFont.c
@@ -105,7 +105,7 @@ static void		DrawCharsInContext(Display *display, Drawable drawable,
  * To avoid an extra copy, a TKNSString object wraps a Tcl_DString with an
  * NSString that uses the DString's buffer as its character buffer.  It can be
  * constructed from a Tcl_DString and it has a DString property that handles
- * converting from an NSString to a Tcl_DString
+ * converting from an NSString to a Tcl_DString.
  */
 
 @implementation TKNSString
@@ -133,7 +133,6 @@ static void		DrawCharsInContext(Display *display, Drawable drawable,
 	_string = [[NSString alloc] initWithString:aString];
 	self.UTF8String = _string.UTF8String;
     }
-    printf("Initialized with string %s\n", self.UTF8String);
     return self;
 }
 
@@ -166,32 +165,16 @@ static void		DrawCharsInContext(Display *display, Drawable drawable,
 	 * The DString has not been initialized. Construct it from
 	 * our string's unicode characters.
 	 */
-
-	char buffer[2*TCL_UTF_MAX];
-	unsigned int index, length, ch;
+	char *p;
+	int index;
 
 	Tcl_DStringInit(&_ds);
-#if TCL_UTF_MAX == 3
-	for (index = 0; index < [_string length]; index++) {
-	    UniChar uni = [_string characterAtIndex: index];
-
-	    if (CFStringIsSurrogateHighCharacter(uni)) {
-		UniChar low = [_string characterAtIndex: ++index];
-		ch = CFStringGetLongCharacterForSurrogatePair(uni, low);
-	    } else {
-		ch = uni;
-	    }
-	    length = TkUniCharToUtf(ch, buffer);
-	    Tcl_DStringAppend(&_ds, buffer, length);
-	}
-#else
+	Tcl_DStringSetLength(&_ds, 3 * [_string length]);
+	p = Tcl_DStringValue(&_ds);
 	for (index = 0; index < [_string length]; index++) {
-	    ch = (int) [_string characterAtIndex: index];
-	    length = Tcl_UniCharToUtf(ch, buffer);
-	    Tcl_DStringAppend(&_ds, buffer, length);
+	    p += Tcl_UniCharToUtf([_string characterAtIndex: index], p);
 	}
-
-#endif
+	Tcl_DStringSetLength(&_ds, p - Tcl_DStringValue(&_ds));
     }
     return _ds;
 }
@@ -1056,7 +1039,7 @@ TkpMeasureCharsInContext(
     [attributedString release];
     [string release];
     length = ceil(width - offset);
-    fit = (Tcl_UtfAtIndex(source, index) - source) - rangeStart;
+    fit = (TkUtfAtIndex(source, index) - source) - rangeStart;
 done:
 #ifdef TK_MAC_DEBUG_FONTS
     TkMacOSXDbgMsg("measure: source=\"%s\" range=\"%.*s\" maxLength=%d "
diff --git a/macosx/tkMacOSXPrivate.h b/macosx/tkMacOSXPrivate.h
index a0645f7..be69fcd 100644
--- a/macosx/tkMacOSXPrivate.h
+++ b/macosx/tkMacOSXPrivate.h
@@ -527,7 +527,7 @@ VISIBILITY_HIDDEN
  * byte sequence as initial data.  So we add a new class which does provide
  * such a constructor.  It also has a DString property which is a DString whose
  * string pointer is a byte sequence encoding the NSString with the current Tk
- * encoding, namely UTF-8 if TCL_MAX >= 4 or CESU-8 if TCL_MAX = 3.
+ * encoding, namely UTF-8 if TCL_UTF_MAX >= 4 or CESU-8 if TCL_UTF_MAX = 3.
  *
  *---------------------------------------------------------------------------
  */