[PATCH] win32: backout 1a9ebc83a74c

[Angel Ezquerra]

On Tue, May 6, 2014 at 7:49 AM, Adrian Buehlmann <adrian at cadifra.com> wrote:
> On 2014-05-06 01:50, Pierre-Yves David wrote:
>> On 05/05/2014 04:02 PM, Steve Borho wrote:
>>> # HG changeset patch
>>> # User Steve Borho <steve at borho.org>
>>> # Date 1399106034 -7200
>>> #      Sat May 03 10:33:54 2014 +0200
>>> # Branch stable
>>> # Node ID 4898c37e03c0b804d6ca8c3ad341dceb53ff5806
>>> # Parent  cadad384c97c7956c98f3c9b92d8cc40fa16d93b
>>> win32: backout 1a9ebc83a74c
>>
>> This have been queued in the clowncopter. Thanks!
>
> Thanks. And - for what's worth - fine be me.
>
> (funny "crew" repo names you guys have nowadays...)
>
>> You can now prepare to advocate for a quick 3.0.1 along side and excuse
>> for not finding this bug during the 3.0-rc period.
>
> Given how small the TortoiseHg team nowadays is and what little
> resources they have, they really don't have much to excuse (Steve in
> particular).

Actually I suspect that I hit this error during the RC period (I also
lost my whole mercurial.ini file at some point). Unfortunately at the
time I was hacking on the TortoiseHg settings code and I just thought
that it was my changes that had caused the problem.

I think the main reason why this bug was not caught earlier is because
there were no TortoiseHg RC builds during the RC period! If there had
been it is quite likely that we would have found the issue much sooner
(I am pretty sure I would have since I tend to use the newest build
TortoiseHg we have).

IMHO the solution to this problem (or at least a big part of the
solution) would be to have an automated build system for the
TortoiseHg installers.

Cheers,

Angel