Re: string change in GTK+

[Matthias Clasen <mclasen redhat com>]

On Tue, 2006-08-15 at 19:48 +0200, Christian Rose wrote:
> On 8/15/06, Matthias Clasen <mclasen redhat com> wrote:
> > I have fixed bug 351063, which leads to a changed string in 2.10.2:
> >
> > <%s> element has neither a \"name\" nor an \"id\" element
> >
> > is factually wrong, and has been replaced by
> >
> > <%s> element has neither a \"name\" nor an \"id\" attribute
> >
> >
> > I have performed the necessary surgery on all po files to keep the
> > old translations.
> 
> Was that really the intention? By doing this surgery and
> unfuzzymarking all translations of this message, the factual error
> will continue to be present in the translations, and more importantly
> you've removed the notification to translators that they should update
> their translations of this message. So the translations will still
> talk about elements, and translators won't notice.
> 
> If you had just left the po files as they are, the message would
> become fuzzymarked, which is the signal to translators that they need
> to update their translations of that message.
> 
> So, unless your intent was really to keep the old, no longer correct
> translations (thus introducing new bugs), I would kindly ask you to
> revert your changes to the po files in this case.

I'll do that after 2.10.2 is out. The intention was to avoid loosing 
an almost correct translation (at least good enough to convey the intent
of the warning) in the imminent release...