>>>>> On Fri, 12 Dec 2003 00:39:34 +0900, Ito Kazumitsu <[EMAIL PROTECTED]> said: kaz> CharToByteEUC_JP exists because it had existed before kaz> CharToByteIconv was born.
kaz> CharToByteIconv may be better than CharToByteEUC_JP but the kaz> benefit of CharToByteEUC_JP is that it can be used without kaz> libiconv. Thanks. I see. How about making CharToByteEUC_JP subclass of CharToByteIconv and use table if it failed to initialize? kaz> If you prefer to use CharToByteIconv, use the encoding name kaz> "EUC-JP" instead of "EUC_JP". But currently EUC-JP/EUCJP are both defined as an alias of EUC_JP by ConverterAlias. I want to use CharToByteIconv, because loadTable of CharToByteEUC_JP is really slow on low-speed CPU (and no-jit) environment (it takes a few MINUTES!). --- Atsushi Nemoto _______________________________________________ kaffe mailing list [EMAIL PROTECTED] http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
