In the current implemenation of some converters, for example, CharToByteISO2022 series and most of subclass of CharToByteDoubleByte, the output byte should be always big enough to hold one character conversion. Say, example of CharToByteISO2022JP, one unicode character is possibly mapped to 5 bytes of sequence. And when output buffer does not hold all the byte to converted one character, it will throw ConverionBufferFullException.

However, if the convert hold the remnant byte to next time convert is called. Even if the output buffer is just one byte, calling convert some time can get the result.

There is also a problem in getMaxBytesPerChar(). Not all the convert does return correct value.

koushi.takahashi@japan 1999-07-09