Re: big5 and gettext

["R.I.P. Deaddog" <maddog linuxhall org>]

On Tue, 22 May 2001, Pablo Saratxaga wrote:

> Kaixo!
>
> Mmh, I'm thinking maybe there is a solution to the big5 encoding problem...
>
> what if the po files are stored in the new format (that is, in real big5);
> and have a small script that checks the version of gettext, and
> eventually, if an old version is present, rewrites the zh_TW*.po file.
>
> The problem is with the '\' that are part of big5 encoding; they always
> are the second byte of a two byte char which first byte is an 8bit one;
> so if there is a perl guru to do a one liner to convert all X\ to X\\
> (with X being any 8bit char) of a po file, then including that script

Actually, 0xa1 <= X <= 0xfe . But does it need to be a guru to write such
one-line script? :D

> could allow the sources to be build on systems with old gettext
> (it will require perl, but it is required for xml-i18n-tools anyway)

Personally I think that this idea is theoretically not bad, but
practically it can be a problem. Asking all software maintainers to
include such fixing can be too much -- working on source to fix only
one locale seems not to have enough merit.

However, I can't think of any better way for now....... -_-

Abel Cheung