Re: git p4 clone not processing branches properly

From: Vitor Antunes <vitor.hda@gmail.com>
Date: Jul 5, 2013, 18:36 UTC
Message-ID: <CAOpHH-WAjjaa3oOXje3u5bBWb=vm-2wG1KYf8oKwH__XLK4R2A@mail.gmail.com>
In-Reply-To: <CAHCaCkLpobqTOUMeK2TP_=VkjRo4P3-dMwt85_CaasSuRNE1Kg@mail.gmail.com>

On Fri, Jul 5, 2013 at 7:11 PM, Matthieu Brucher <matthieu.brucher@gmail.com> wrote:

Show 8 quoted lines

>> Hi Matthieu,
>>
>> Could you please try using //Depot/Project<at>all instead of selecting a
>> specific revision?
>
> I can try. Indeed, at this revision, the two other branches do not yet
> exist. But @all will get everything? Last time, I only got head
> (IIRC).

Our P4 server has a limitation on the number of lines returned by "p4 changes" command, so sometimes I have to use @change_start,@change_stop instead of @all. You might want to use this range limitation to test git-p4 by limiting to a small number of changelists that allows you to check if at least one branch is correctly detected.

Show 7 quoted lines

>> Also, by using that command it means that the following depot paths must
>> exist:
>> //Depot/Project/Branch/Main
>> //Depot/Project/Releases/2013
>> //Depot/Project/Branch/Feature1
>
> Yes, they indeed do.

In this case the problem should not be in branchList configuration.

Show 6 quoted lines

>> I've never used the --use-client-spec, so I'm not sure if that will not
>> break the branch detection code.
>
> I need to do that because if I don't, the depot is clobbed with
> binaries. Or perhaps if I put some .gitignore stuff, I might not do
> this?

Keep using it, at least for now. If everything else fails we can look at this again.

Cheers, Vitor