moving all .po files to UTF-8

[Keld Jørn Simonsen]

On Sat, Mar 03, 2007 at 01:15:48AM +0000, Dimitris Glezos wrote:
> O/H Keld Jørn Simonsen ????????????:
> > On Fri, Mar 02, 2007 at 01:24:53AM +0000, Dimitris Glezos wrote:
> >>>
> >>> Rather than doing it at build time, for all the software that's translated
> >>> in/by Fedora, it seems simplest to do it in the .po files itself. Therefore,
> >>> at some point in the next few days, we intend to do this in the CVS
> >>> repository.
> >>>
> >>> Does anyone forsee any issues with this?
> > 
> > I am not happy about this, and in general I think it is the wrong way to
> > go. Eg. all my own tools for translation are using other charsets than
> > UTF-8
> 
> Hi Keld,
> 
> I think most people would agree that UTF-8 is the "right" way to go -- that's
> the reason for its existence, right? It's something that can make sure
> everything will work for all languages.

maybe most people, but this is not the question, - most people already
have gone to UTF-8 already. 

> I'm thinking that this is an opportunity for all who don't use it in their tools
> to start using it. :)

It would be a backwards step for me, UTF-8 allows some characters that I
automatically avoid using 8859-15. Thus my translations would have
fewer errors not using utf-8. And my tools run 8859-15.

> >, and also, if the best charset for the source, according to a
> > specific translator, would be something else than utf-8, and cvs is a
> > source control system, then the cvs repository should store the source,
> > and not something derived from the source.
> 
> Not sure I understand what you mean. Since UTF is a superset of all charsets
> then it's at least equally good to those in representing characters. Most of our
> tools not only support it, but use it as a default.

I am talking about what a source control system is. 

yes, your tools use utf-8, my tools run 8859-15. Lets allow both you and
me to work for the fedora project.

Best regards
keld