Re: [PATCH/RFC 0/7] i18n: mark missing C messages

[Junio C Hamano <gitster@pobox.com>]

Ævar Arnfjörð Bjarmason  <avarab@gmail.com> writes:

> This is an RFC because the first patch still leaves one remaining
> gitdir test failing under GETTEXT_POISON, and I won't be able to find
> out why tonight. Help welcome with figuring out why (in which case
> this should all be OK), you just need too build with
> GETTEXT_POISON=YesPlease and run t/t0001-init.sh.

I didn't even compile test this one, but I highly suspect that all the
failures you saw come from the fact that .git file your patch produces
with gettext poisoning (or any l10n) would be malformed and cannot be
parsed by read_gitfile_gently() anymore.

I took a look at the remainder of the series, and except for the "do not
call what you know is a singleton 'them'" grammo, everything looked fine
to me.

I am tempted to merge them to 'master' before rc2; the alternative is to
queue it on ab/i18n-fixup and keep it in 'next' until the 1.7.5 final, and
have other people build on top of ab/i18n-fixup topic.