public final CoderResult encode(CharBuffer in,
Considering that there is a supplemental character which represents by 4 bytes like "\ud800\udc00", and it was divided into two input buffers. e.g. "\ud800" is at the end of input1 and the "\udc00" is at the begining of input2. when invoking the method encode(input1,out,false), it deals with "\ud800" as a malformed charactor and I think it should be reversed for next invocation like encode(input2,out,true).
And it aslo must be handled correspondingly in method public final ByteBuffer encode(CharBuffer in).
===Test segment to reproduce the bug (fail on RI 5.0 and Harmony. I think it's also a bug for RI 5.0)
CharsetEncoder encoder = Charset.forName("utf-8").newEncoder();
CharBuffer in1 = CharBuffer.wrap("\ud800");
CharBuffer in2 = CharBuffer.wrap("\udc00");
ByteBuffer out = ByteBuffer.allocate(4);
CoderResult result = encoder.encode(in1, out, false);
result = encoder.encode(in2, out, true);