Charset.forName("EUC-TW").newEncoder().canEncode(c)
returns false for all c <= 0xff.
But many of those characters (in particular, all the ASCII characters),
can be encoded.

(This is a part of the fix for 5065777)

###@###.### 2004-06-22