Re: Please check your sources for strings not marked for translation before release

[Kenneth Nielsen <k nielsen81 gmail com>]

Den 07-05-2012 14:04, Jorge González skrev:
> Hello,
>
> On Mon, May 7, 2012 at 1:59 PM, Kenneth Nielsen<k nielsen81 gmail com>  wrote:
> > Hallo developers.
> >
> > Since the release of GNOME 3.4 we have received a steady stream of emails
> > saying "This is not a string freeze break, we just forgot to mark the
> > strings for translation". So much so, that the statistics e.g. for my
> > language (which has not been touched since release) now counts 63*
> > untranslated or fuzzy strings in a source that was at 0 at release time,
> > three weeks ago.
> >
> > I understand the reason for not counting marking existing strings as a
> > string freeze break. But please understand that even though you are in your
> > right to fix things like this post release, and even though it off course
> > makes sense to do it, it does not mean that it is not annoying and a burden
> > for translators.
> >
> > The reason for this is that we concentrate large amounts of work in the
> > period just before release (where string changes are relatively quiet), so
> > therefore it is also an advantage for us if we can finish as much of it as
> > possible in that period (of course also because a lot of us has an extra hat
> > as distribution translators for distributions that follow within a month or
> > so of the gnome release). On top of that, small updates are
> > uncharacteristically expensive in a work/string-sense because they involve
> > the same amount of emails back and forth for proofreading and git work.
> >
> > So please... If you could make a bit more of an effort of checking your
> > sources for non-internationalized strings before release that would be
> > great. IMHO 63 is just a bit on the high side of where this number should be
> Can't this be half-automatized? like with a bash script that checks
> the strings, which are marked and which not?
> It could be helpful.

I guess it may depend on the syntax of the language. But I guess for 
most languages it should be doable[1] to make a script the output all 
the strings that has not been marked. It might of course contain false 
positives in the form of strings purposely not marked for translation, 
but I guess it should be easier to review this output than the entire 
source.

The pseudo translation approach as mentioned by Friedel is of course 
also an options, I'm just not sure if it is as easy for developers to 
"exercise the application" as suggested in the wiki page.

Regards Kenneth

[1] By someone with an appropriate level of reg-exp-fu

> > Regards Kenneth
> Kind regards.
>
> > * These 63 off course also include genuinely new strings due e.g. to bug
> > fixes.
> > _______________________________________________
> > gnome-i18n mailing list
> > gnome-i18n gnome org
> > http://mail.gnome.org/mailman/listinfo/gnome-i18n