[petsc-dev] Pause-for-approval Pipelines?

Barry Smith bsmith at petsc.dev
Thu Aug 27 17:34:15 CDT 2020

> On Aug 27, 2020, at 4:55 PM, Jed Brown <jed at jedbrown.org> wrote:
> Barry Smith <bsmith at petsc.dev> writes:
>>> - readdocs pipeline
>>> - merge-pipeline - auto starts - does the pre stage and pauses.
>> how are they related? 
> Not at all.   The ReadTheDocs pipeline is managed by readthedocs.org.
>> I understood Jed wanted the readdocs pipeline done automatically so it was easy to review any readthedocs changes since the properly formatted new docs with the changes was made automatically in the MR and pushes. I am fine with this, though it seems silly to always do it instead of not just doing it when there is a change to docs or maybe a docs label. Just curious, is this done with a merge to master ?
> We don't have control at that granularity.  We can disable the feature, but then we don't have a good way to review changes there.  I want to make writing and reviewing documentation changes as low friction as possible (because our docs, especially the manual, are pretty low quality and often out of date).
> I figure if ReadTheDocs feels like they're wasting too much compute, they can add a finer grained control.

As I said I am fine with this readthedocs, assume it works.
>> The second pipeline I don't understand. Aside from letting people who don't use the GitLab API to  not have to hunt for the Pipelines window and cut and past their branch name in and the merge to master business does it serve any other purpose? Could we unset this for certain people who don't like it in the .gitlab-ci.yaml file?

  If the second pipeline has nothing to do with readthedocs then why did your MR also include it? Was it two completely different  unrelated things in one MR? Or does readthedocs require MR pipelines to be on and somehow this requires a regular MR pipeline be run even though not directly connected to readthedocs, your stopping it and waiting is the only way to prevent it from running fully? What about for some people supporting stopping it and cancelling it once readthedocs is done or even deleting it then so it doesn't clog the pipeline window with irrelevant stuff?

   I understand the utility of this second pipeline and the button for many people but it does clog up the pipelines. I think some people should be able to turn it off.

> What's wrong with using the API to release the paused job instead of using it to start a fresh pipeline?

  Generally I like to pass the Pipeline before making a PR. So the test on creating a new MR is annoying. Yes after the initial MR I might be able to release the paused job in lieu of starting pipelines fresh. It would be nice to send some pushes that don't trigger a pipeline start at all because I know I don't need one. Maybe that is possible, I'll need to investigate.

More information about the petsc-dev mailing list