Re: What's cooking in git.git (Nov 2018, #03; Wed, 7)

From: Junio C Hamano <gitster@pobox.com>
Date: Nov 7, 2018, 22:09 UTC
Message-ID: <xmqqpnvgbkvd.fsf@gitster-ct.c.googlers.com>
In-Reply-To: <CACsJy8CuYgX9yFisJerf9C_6-hyyZD8DjzTsQ7v-uKpbm5LvQQ@mail.gmail.com>

Duy Nguyen <pclouds@gmail.com> writes:

Show 20 quoted lines

> On Wed, Nov 7, 2018 at 2:09 PM SZEDER Gábor <szeder.dev@gmail.com> wrote:
>>
>> On Wed, Nov 07, 2018 at 06:41:45PM +0900, Junio C Hamano wrote:
>> > * nd/i18n (2018-11-06) 16 commits
>> >  - fsck: mark strings for translation
>> >  - fsck: reduce word legos to help i18n
>> > ...
>> >  More _("i18n") markings.
>>
>> When this patch is merged into 'pu' all four tests added to
>> 't1450-fsck.sh' in b29759d89a (fsck: check HEAD and reflog from other
>> worktrees, 2018-10-21) as part of 'nd/per-worktree-ref-iteration'
>> below fail when run with GETTEXT_POISON=y.  The test suite passes in
>> both of these topics on their own, even with GETTEXT_POISON, it's
>> their merge that is somehow problematic.
>
> Not surprising. The i18n series makes fsck output localized strings
> and without updating grep to test_i18ngrep, new tests will fail. If
> 'pu' was passing before, I'm ok with just ejecting this series for
> now. Then I wait for the other to land, rebase, fixup and resubmit.

Let me first see if I can come up with a merge-fix that can be carried around during the time this topic cooks, before talking about dropping and reattempting the series.

For a change like this, a time-window in which the codebase is quiescent enough may never come, and because the changes go all over the place, mostly but not entirely repetitive, it costs a lot, not just to write but also to review them.