Tomohiro KUBOTA wrote: > I imagine Text::Iconv is not UTF-8-aware at all, because it divides > a multibyte character (UTF-8 is multibyte; most of ISO-8859 characters > are two bytes and east Asian characters are three bytes) into two parts. It is as utf-8 aware as is iconv(), but it does not use the same internal representation perl does for unicode. I think. > I think that usage of Encode:: module will work fine, but I have some > questions. > > debconf-1.2.x uses libtext-iconv-perl already. Why don't we use this? > > Encoding:: module is separated into a few sub-modules, like Encoding::JP > and so on. This may cause complex implementation or inability of > support of some encodings. locale -m lists 196 supported encodings on my system, while: joey@dragon:~>perl -le 'use Encode; use Encode::CN; use Encode::JP; use Encode::KR; use Encode::TW;print join("\n", Encode->encodings())' | wc -l 32 The other reason I didn't use it is I am worried that there might be problems if a translaton is converted: translator's encoding --------> utf-8 -------------> display encoding iconv perl Encode -- see shy jo
Attachment:
pgpxK4j5_bI_E.pgp
Description: PGP signature