Patch [2/5] not part of the rest of the series

[Stephen Finucane]

On Thu, 2020-03-19 at 10:21 -0400, Konstantin Ryabitsev wrote:
> Hello:
> 
> I have a rather strange case here:
> 
> https://patchwork.kernel.org/project/linux-rdma/list/
> 
> If you look for "258273" in the page, you will find an oddball situation 
> where a [v2,2/5] is randomly recorded as part of "Untitled series", even 
> though the rest of the patches in that series are processed properly.
> 
> I looked through the headers and all the threading appears to be 
> correct. It would be easy to sketch this up to random failure, but the 
> exact same thing happened with the v1 of that series:
> 
> https://patchwork.kernel.org/patch/11442933/
> https://patchwork.kernel.org/patch/11442939/
> 
> Again, it's 2/5 that is oddly out of place.
> 
> Any guesses?

You're looking for #241 [1]. tl;dr: We race when creating a series and
the database constraints that should have protected us were not
correct. That's been fixed on master [2] and should form part of
v2.2.0.

Stephen

[1] https://github.com/getpatchwork/patchwork/issues/241
[2] https://github.com/getpatchwork/patchwork/commit/9f72eb793dfb6e9d7ff54465d4b07291e9a75e38