Re: I18N: automatic determination of potfiles list

[philippt at t-online dot de (Philipp Thomas)]

* Zack Weinberg [Mon, 29 Oct 2001 21:17:09 -0800]:

>This patch causes us to calculate the list of files to be scanned for
>message translation on the fly.

To start with, thanks for the work you've done.

>Finally, the baroque po/Makefile.in.in has been scrapped and a minimal
>set of rules inserted into gcc/Makefile.in.

Agreed, as with POTFILES gone, there isn't any need for it.

>Assuming this patch is acceptable, I would like to set up automatic
>updating of gcc.pot through the snapshot script or some other
>mechanism. The file currently in the tree is perennially out of date.

Yes, that would indeed be preferable. Up til now the updating had to be done
manually and thus lagged when I forgot to do so.

>[It may even make sense to run msgmerge automatically against the
>translations that we do have; I'm not sure what the issues are there.]

No, that doesn't make much sense as msgmerge sometimes merges messages
incorrectly. In that case you would get rather strange results. Given that you
have to understand the language, this is best left to the translators and not
done automatically.

But if gcc.pot is updated automatically, we should make it available to the
translation project robot to in turn have the translation teams be notified of a
new version.

>Related notes: At this time xgettext does not know how to parse Ada and
>therefore the Ada front end cannot have its messages translated.  I'm
>not sure what should be done about that.

IMO xgettext should be made to understand Ada. I'll ask Bruno Haible
<haible@ilog.fr>, the current comaintainer of the gettext tools what would be
needed to support Ada.

Philipp

-- 
The three enemies of the hacker?
Sunlight, fresh air and the unbearable roaring of the birds.

-- 
Penguins to save the dinosaurs
                     -- Handelsblatt on Linux for S/390