Re: Further protections on master branch

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Sep 5, 2018 at 3:56 PM Alfredo Deza <adeza@xxxxxxxxxx> wrote:
>
> On Wed, Sep 5, 2018 at 10:41 AM, Gregory Farnum <gfarnum@xxxxxxxxxx> wrote:
> > Hey all,
> > After a discussion on options in the leadership team today, I ticked a
> > box in the Github interface that prevents repository administrators
> > from merging PRs to master that don't pass status checks (ie, make
> > check). Hopefully this was rare anyway, and what we've really done is
> > added some security against accidental "git push" of test branches
> > that aren't ready yet and being just that little bit more sure that we
> > aren't breaking things for other people.
>
> There is exactly 0 gain here for a project like ceph-volume which
> doesn't benefit from (nor it can't affect) make check.
>
> We knew this when we started ceph-volume out of tree, but now that we
> are in-tree, it is certainly a lot of trouble that we avoid by always
> merging
> regardless of what make check is reporting, not to mention that it
> takes about 40 min. to complete (ceph-volume functional tests take
> half the time).
>
> I understand this may be for the greater good of the project as a
> whole, but I have to raise this as a problem for components like
> ceph-volume that are undermined by these
> changes

This isn't really unique to ceph-volume.  We could equally point out
that RADOS gets held up if make check is broken with a something from
rbd/rgw/cephfs.

The underlying benefit here isn't about how we merge PRs, it's about
having the tip of the master branch pass its own tests.  That benefits
everyone in the Ceph ecosystem, even out of tree components that might
be continuously integrated against Ceph's master branch.

The benefit to you personally is that once we have motivated everyone
else to stabilize make check, you get a reliable green tick for your
ceph-volume PRs, rather than having them interrupted with annoying red
crosses.

John

> >
> > But maybe this is going to be an issue for us – if you find yourself
> > needing to override, you have the ability to turn it off
> > (temporarily?) — please let me know if it causes you trouble! If it
> > goes well I'll consult with the backports team and similarly check
> > that box for our stable branches.
> > -Greg



[Index of Archives]     [CEPH Users]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux