Re: Broken gfxterm_menu tests

[Colin Watson]

On Mon, Nov 18, 2013 at 06:00:59PM +0000, Colin Watson wrote:
> On Mon, Nov 18, 2013 at 05:10:13PM +0100, Vladimir 'φ-coder/phcoder' 
> Serbinenko wrote:
> > Colin Watson wrote:
> > > Is this specific to my environment, or is it happening for everyone?  In
> > > the latter case I guess the checksums ought to be updated.
> > 
> > Judging by your other commits you don't have locales installed. The
> > tests you point check for gettext among other things.
> 
> I certainly have locales installed in general, but since I'm running
> from a clean git checkout I don't have po/*.po in place.  Are you saying
> that that would cause this failure?
> 
> I think our tests should pass when run from a clean git checkout,
> without needing to fetch translations from an external source.

For the life of me I cannot get this to pass.  I tried running
./linguas.sh first, and patching grub-shell to pass --locale-directory
to grub-mkrescue so that it's definitely using the PO files from the
source directory; no luck as yet.  Any help would be greatly
appreciated.


That said: I propose that we should occasionally run ./linguas.sh and
commit the result to git.  While these files are autogenerated from our
point of view and our copies wouldn't be the furthest-upstream versions,
they clearly aren't entirely machine-generated so I think it is
reasonable to store them in revision control.  This would have at least
the following advantages:

 * Our build processes would no longer be vulnerable to an external
   server potentially going down for an extended period of time; we'd be
   stuck with outdated translations until somebody fixed it or came up
   with a workaround, of course, but that usually isn't fatal.

 * It would be easier to manage branches of stable releases, rather than
   assuming that translations downloaded for trunk will match the POT
   files for a stable release.

 * Tests would be able to pass from a clean git checkout without relying
   on an external server, improving QA reliability.

 * It would be easier to make and test branches while offline.

 * The translations shipped with a release tarball could be tagged in
   git so that it's easy to investigate bugs in them.

 * Downstream distributors would be able to use git branches without
   having to fill in additional files.  (This has been a low-level
   annoyance for me for some time.)

What do you think?

-- 
Colin Watson                                       address@hidden