Tk_MeasureChars: Use Tcl_UtfToUnichar() for lax UTF-8 parsing

instead of strict parsing with FcUtf8ToUcs4() [fix/workaround for #1185640]

1 files changed, 7 insertions, 7 deletions

diff --git a/unix/tkUnixRFont.c b/unix/tkUnixRFont.c
index 2557ab0..cd4a4f0 100644
--- a/unix/tkUnixRFont.c
+++ b/unix/tkUnixRFont.c
@@ -8,7 +8,7 @@
  * See the file "license.terms" for information on usage and redistribution
  * of this file, and for a DISCLAIMER OF ALL WARRANTIES.
  *
- * RCS: @(#) $Id: tkUnixRFont.c,v 1.8 2004/11/28 19:00:49 jenglish Exp $
+ * RCS: @(#) $Id: tkUnixRFont.c,v 1.9 2005/04/19 05:50:06 jenglish Exp $
  */
 
 #include "tkUnixInt.h"
@@ -511,15 +511,15 @@ Tk_MeasureChars(tkfont, source, numBytes, maxLength, flags, lengthPtr)
     curByte = 0;
     sawNonSpace = 0;
     while (numBytes > 0) {
-	clen = FcUtf8ToUcs4((FcChar8 *) source, &c, numBytes);
+	Tcl_UniChar unichar;
+
+	clen = Tcl_UtfToUniChar(source, &unichar);
+	c = (FcChar32)unichar;
 
 	if (clen <= 0) {
-	    /*
-	     * This should not happen, but it can, due to bugs in Tcl
-	     * (i.e., [encoding convertfrom identity]).
-	     */
+	    /* This can't happen (but see #1185640) */
 	    *lengthPtr = curX;
-	    return ++curByte;
+	    return curByte;
 	}
 
 	source += clen;