From: Luke Diamand <luke@diamand.org>
Date: Thu, 23 Apr 2015 10:03:02 GMT
Subject: Re: [PATCH V3 0/2] git-p4: improve client path detection when branches are used
Message-ID: <CAE5ih79Ju6D0mZ75Z1UH2n2O6s9AWuxNgpojr=vTggEYY=7E3w@mail.gmail.com>
In-Reply-To: <9A7A7CAA-66D5-48B4-83E7-267FDDA7C66A@gmail.com>

On 23 April 2015 at 09:37, Vitor Antunes <vitor.hda@gmail.com> wrote:

> That was a good combination to test. In fact, I am using such
> a client spec at my work place to exclude the import from
> Perforce of a folder that only contains binary files, but I never
> even considered to add files to that folder from git!
> Although I do agree that git-p4 should be able to exit sanely
> in this scenario, it is also my opinion that this is a different
> scenario from the one I'm tryig to fix in this set of patches and
> that it should not be enough to stop this merge.
>
> I will take this scenario into consideration, create a new test
> case and finally fix git-p4 to exit sanely in such a scenario.
> This new test will also be able to show that folder exclusion
> is working perfectly during import, which is important to
> guarantee that that functionality is not broken in future.
>
> BTW, no kudos is necessary because I've already walked this
> path before :) I've introduced branchList and improved how
> git-p4 looks for the original changelist used to create the new
> branch in Perforce side. If I remember correctly, many of the
> test cases in 9801 file were also created by me before Pete
> started splitting the git-p4 test file into topics.
>

Yes, I think you're right that this is a different, albeit related
problem. It was broken before, and it's still broken. So I'm happy
with this change - Ack.

Teaching git-p4 to handle this corner case would be a good thing to do
as a separate followup if you're OK to do that.

Thanks!
Luke