Charsets which are prefixed by a BOM need to include the size of it into maxBytesPerChar() to fulfil bug JDK-8230531.

The downside is, that the buffer, calculated from maxBytesPerChar(), is near twice in size than needed.

With this new API, the actually needed buffer size could be calculated with:

    int bufSize = encoder.minBytesPerSequence() + (encoder.maxBytesPerChar() - encoder.minBytesPerSequence()) * sequenceLength;
    // or
    int bufSize = encoder.maxBytesPerChar() * sequenceLength - encoder.minBytesPerSequence() * (sequenceLength - 1);