[PATCH 1 of 1 STABLE] dirstate: don't write repo.currenttransaction to repo.dirstate if repo is None (issue4983)

[Sietse Brouwer]

> Thanks, you patch had some bad damage from the your email client. Consider
> using the patchbomb extensions or the pushgate to send a patch next week.

Noted. I thought sending as plain text would keep Gmail from mangling
it, but nooo. Have updated the wiki to make that clear. Excited to try
out the pushgate in future. Thank you very much for the repair work.

https://www.mercurial-scm.org/wiki/ContributingChanges#Experimental_submission_via_push_gate

>> Matt's e-mail backlog is terrifying

> Augie is the goaltender this week, so the important backlog is the patchwork one.

That's good news. Pity about missing 3.6.2, yes... ah well, such is life.

Cheers,
Sietse

On 3 December 2015 at 10:20, Pierre-Yves David
<pierre-yves.david at ens-lyon.org> wrote:
>
>
> On 12/02/2015 05:22 PM, Sietse Brouwer wrote:
>>
>> This fixes issue4983 [1]: post-init and post-clone hooks always crash
>> (after hg init / hg clone)
>
>
> Thanks, you patch had some bad damage from the your email client. Consider
> using the patchbomb extensions or the pushgate to send a patch next week.
>
> I've pushed this to the clowncopter.
>
>> I realise that Matt's e-mail backlog is terrifying right now [2]; but
>> the bug is a regression, and the fix is a simple one, so I thought I'd
>> send it in anyway.
>
>
> Augie is the goaltender this week, so the important backlog is the patchwork
> one.
>
> The regression is indeed pretty bad. (and the scheduled bugfixe version was
> 2 day ago…)
>
> --
> Pierre-Yves David