[TCLCORE] Re: Translations strings in CVS

[Giorgos P. <pe...@ii...>]

In Ðåì, 06 Éïýë 2000, lau...@uf... wrote:
> On  6 Jul, Giorgos Petasis wrote:
> > Hi Laurent,
> > 
> > I have done the greek translation and I have attached the file.
> > The file is encoded in utf-8 encoding, as you suggested.
> > 
> > On equestion remains though: Have you patched message catalogs
> > to read things in utf-8 encoding by default?
> > Because this file if simply sourced won't work.
> > Have i done anything wrong?
> > 
> 
> I don't know. Like I said before, I've never touched anythig other than ISO
> 8859-1 (latin 1). Can you resend the file zippped (or gzipped, whatever) and
> copy tc...@aj...? My MUA seems to think it's a text file but
> that may mess things up when I save.
> 
> I'll try to test the loading and stuff later.

Ok, I have gzipped the file and attached it to this mail.
The message file is in utf-8 encoding.

My opinion is that the whole message catalog
facility has to store its messages only in this format. I see no point in
doing a simple source of the file that contains the messages and leave
the code inside this file to do the necessary conversions.
Instead if we decide that files are sourced with a default encoding (a feature which
I think will be in next release, source -encoding), then there will be no need for
conversions. In order to use the message catalog as is right now, I have to place
a "encoding convertfrom iso8859-7..." in front of every (!) string.
And I am not sure if this would not depend on the system encoding...

George