Re: [PATCH v2 6/7] Fix tests under GETTEXT_POISON on git-remote

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 28, 2012, 01:27 UTC
Message-ID: <7vehmrpzzj.fsf@alter.siamese.dyndns.org>
In-Reply-To: <CANYiYbGudJ56j=o5SDJur2dv2=icCe2FYad88dV_5tBzLAbZnA@mail.gmail.com>

Jiang Xin <worldhello.net@gmail.com> writes:

Show 16 quoted lines

>> This couldn't have possibly passed with the trailing &&, or am I
>> missing something?  There is already "add another remote" before
>> this test that adds "second" remote.  Is this test about "add
>> yet another remote", or is it checking the result of adding "second"
>> that was done in the previous step?
>
> The trailing "&&“ is a copy & paste error. I only run my fixup in
> GIT_GETTEXT_POISON mode in a harry, not noticed the bypassed
> testcase has this serious bug.
>
> I split the original "add another remote" into two blocks. One is a
> normal testcase, and another has a C_LOCALE_OUTPUT prereq
> flag. This is because other testcases depend on the operations in
> "add another remote" testcase ('git remote add -f second ../two'),
> and these testcases would fail if add C_LOCALE_OUTPUT
> prereq to the whole "add another remote" testcase.

OK, so in short, I think the fix-up I made while queuing (and the typofix in the commit log message for the other one) this round to 'pu' should match what you wanted to do. If you want to make further changes, I wouldn't mind re-queuing an update, but otherwise there is no need to resend this series (and it will be less work for me ;-).

Thanks.