Apply Patchseries Script

David Marchand david.marchand at redhat.com
Thu Sep 28 14:31:44 CEST 2023


On Thu, Sep 28, 2023 at 2:05 PM Aaron Conole <aconole at redhat.com> wrote:
>
> Patrick Robb <probb at iol.unh.edu> writes:
>
> > On Wed, Sep 27, 2023 at 4:22 PM Thomas Monjalon <thomas at monjalon.net> wrote:
> >
> >  27/09/2023 18:31, Patrick Robb:
> >
> >  > 2. Do not apply and run if the series is an RFC series
> >
> >  Not sure about this requirement.
> >  What is the problem in running tests on RFC?
> >
> > I see that currently ovsrobot and UNH Lab have rules saying don't test on RFC series, and Loongson and Intel do test on
> > RFC series. I'm guessing the thinking was something like "RFC patches are at least one stage away from merge, and
> > probably do not represent the final state of the patch, so CI testing is not very valuable." On the other hand, I'm sure in
> > many cases getting that early feedback, even for an RFC, is helpful to developers. I'll bring it up in the CI testing
> > meeting tomorrow and see if any of the CI testing people have an opinion. Anyways, I think all labs should have the
> > same policy, be it testing or not testing on RFC patches.
>
> We do currently skip running RFCs as well.  IIRC they were eating into
> our timing budget on Travis, and we never bothered to re-evaluate after
> the switch to github actions.  I think it would be good to discuss it.

I once missed some issues in a RFC of mine that I only discovered when
sending the first non-RFC patch.

As far as time budget is concerned now, I don't think testing RFC is
that much of an issue.
We have way more people sending 3-4 revisions in a single day than
people sending RFCs :-).


-- 
David Marchand



More information about the ci mailing list