Linux uses a variant of the EUC-JP encoding that only supports the characters from JIS X 0201 and JIS X 0208, not the ones from JIS X 0212. Various parts of the system (in particular the text widgets used by AWT text components) fail when confronted with the three-byte sequences that encode JIS X 0212 characters in EUC. To work around this problem, we need variants of the EUC_JP converters that only support JIS X 0201 and JIS X 0212 (call them CharToByteEUC_JP_LINUX and ByteToCharEUC_JP_LINUX), and use this variant encoding as the default file encoding for the Japanese locales on Linux.