From: Junio C Hamano <gitster@pobox.com>
Date: Tue, 29 Jun 2010 17:22:24 GMT
Subject: Re: [PATCH] Update t3508 status
Message-ID: <7vd3v9ah4f.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20100629165032.GA7699@burratino>

Jonathan Nieder <jrnieder@gmail.com> writes:

> Junio C Hamano wrote:
>> newren@gmail.com writes:
>
>>> Test #8 of t3508-cherry-pick-many-commits.sh
> [...]
>> Curious.
>> 
>> This test is broken in the second parent of 3af0bba8b (i.e. the tip of
>> cc/cherry-pick-stdin), but in 3af0bba8b, i.e. the merge result, it does
>> seem to be fine.
>> 
>> What am I missing???
>
> The fix is in commit v1.7.2-rc0~6^2 (DWIM 'git show -5' to 'git show
> --do-walk -5', 2010-06-01):
>
>  disable --no-walk in this case so ‘git show’ and future ‘git
>  cherry-pick’ can behave as expected.

Ah, that's it.  3af0bba (Merge branch 'cc/cherry-pick-stdin' into next,
2010-06-27) is on 'next', so this has to be queued as an evil merge of
your "DWIM" thing into cc/cherry-pick-stdin.  That way, when the latter
graduates to 'master', the test will be marked as "expect_success".

Thanks.