Michael Sobolev <[EMAIL PROTECTED]> writes: > On Thu, Sep 30, 1999 at 11:24:34PM +0900, Changwoo Ryu wrote: > > Ulrich Drepper, the glibc maintainer, once said he would enhance > > gettext so that gettext() convert the codes on the fly. This is > > possible because there is an encoding information in the catalog > > files: > > > > $ LANG=ru_RU gettext tar "" > > Project-Id-Version: tar 1.13.7 > > POT-Creation-Date: 1999-08-23 03:56-0700 > > PO-Revision-Date: 1999-08-23 09:31+08:00 > > Last-Translator: Const Kaplinsky <[EMAIL PROTECTED]> > > Language-Team: Russian <[EMAIL PROTECTED]> > > MIME-Version: 1.0 > > Content-Type: text/plain; charset=koi8-r > > Content-Transfer-Encoding: 8bit > > $ > Such an enhancement would be nice. But at the moment, this is not > implemented, > is it? So the messages are to go where they should be. From other hand, not > all message catalogs contain this information. What I want to emphasize is > that there is no policy at all with respect to what is mandatory in message > catalogs.
No, it's not implemented. But in logical the gettext catalogs are charset independent. Man pages or catgets catalogs are not. Well, Korean uses only EUC-KR (or, at least, 99.99%). But if I live in a country in which I should use more than 2 encodings, I would urge Ulrich rather than preparing multiple catalog files. :) -- Changwoo Ryu
