[hibernate-dev] Not letting Jenkins auto-close PRs upon build failures

[Gunnar Morling]

2016-01-08 14:39 GMT+01:00 Sanne Grinovero <sanne at hibernate.org>:
> Do you have an example of such a PR being closed?

Yes, https://github.com/hibernate/hibernate-ogm/pull/610 whose
auto-closing half a dozen times nearly drove me nuts ;)

>
> For sure that was never our intention, but be aware that such options
> can be set globally on the ci server as project detaults, but can be
> overriden in a specific build task. So I'd check the build task which
> did this as for example Search PRs seem fine.

It's the OGM PR builder, maybe you could check this one?

>
> In my experience every time we upgrade Jenkins plugins there are such
> issues like changes of defaults, or sometimes the plugin configuration
> formats change so we lose some details.
>
> I didn't update plugins recently, but a service crash - AFAIR it was
> the 24th of December - had me restart the server. At that point I also
> installed system (security) patches, but while this might have
> triggered a micro update of Jenkins itself for some security reason,
> it shouldn't have affected plugins.
> Another option I can think of is that someone had previously requested
> the update, but didn't reboot the service so the update was only
> applied during my reboot.
>
>
>
>
> On 5 January 2016 at 13:02, Emmanuel Bernard <emmanuel at hibernate.org> wrote:
>> +1, but I don’t know how ;)
>>
>>> On 05 Jan 2016, at 13:55, Gunnar Morling <gunnar at hibernate.org> wrote:
>>>
>>> Hi,
>>>
>>> Not sure whether that's a new behaviour, but it's super annoying that
>>> the pull request builder automatically closes pull requests if there
>>> is a build failure. Can we change this so the PR stays open, please?
>>>
>>> The current behaviour is highly frustrating, you need to re-open the
>>> PR again after fixes, and in the worst case a PR falls through the
>>> cracks because it's moved out to the "Closed" tab right away. It's
>>> especially bad for contributor PRs where the failure is caused by
>>> ourselves, e.g. build environment issues.
>>>
>>> Thanks,
>>>
>>> --Gunnar
>>> _______________________________________________
>>> hibernate-dev mailing list
>>> hibernate-dev at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>>
>>
>> _______________________________________________
>> hibernate-dev mailing list
>> hibernate-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/hibernate-dev
>
> _______________________________________________
> hibernate-dev mailing list
> hibernate-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hibernate-dev