Hello guys,
I just wanted to announce that the service is already deployed and
that I am trying to figure out all the quirks and ask you to please be
patient.
But you were faster :-)
Thank you for the reports, please ping me with anything more that you
find. I am also trying to monitor what is happening.
The issue that Miro found
https://bugzilla.redhat.com/show_bug.cgi?id=2133080#c19
was caused by an URL-encoded caret symbol in the package name, which
is IMHO a Copr bug. I already sent a PR
https://github.com/fedora-copr/copr/pull/2454
The issue that Frank found
https://bugzilla.redhat.com/show_bug.cgi?id=2119983
I am not sure, looks like a temporary unavailability of the URL
because when I resubmitted the package, it built successfully. Not
sure what should be done about this, I will probably wait if this
happens again.
Jakub
On Thu, Jan 5, 2023 at 1:45 AM Frank Crawford <frank(a)crawford.emu.id.au> wrote:
On Wed, 2023-01-04 at 23:41 +0100, Miro Hrončok wrote:
> On 02. 12. 22 18:20, Jakub Kadlcik wrote:
> > Hello folks,
> >
> > for a couple of years now, I've been interested in the Fedora
> > package
> > review process. Our queue is in hundreds of waiting packages for as
> > long as I can remember. I believe the situation can be improved.
> >
> > Since summer, I did ~40 package reviews to get a better grasp of
> > the
> > situation and realized what I want to do.
> >
> > I started working on a service [2] that listens to fedora-messaging
> > and for every new RHBZ review ticket or a new comment with updated
> > packages, it submits a build in Copr. Thanks to this [1] feature,
> > Copr
> > automatically runs the fedora-review tool and generates the
> > review.txt file. Once the build is finished, my service gets the
> > message and generates a helpful comment (so far only to STDOUT).
> >
> > **Unless there is general disapproval, I am planning to let it post
> > the comments to Bugzilla.**
> >
> > So far the benefit is limited but, it still can immediately tell
> > the
> > contributor that their package is broken and fails to build, and
> > also
> > saves a reviewer the time of running the fedora-review tool
> > manually. However, I also implemented support for the fedora-review
> > tool to generate a JSON output next to the standard review.txt. The
> > PR
> > is still pending [2], please review it if you can. Once this is
> > released, I can parse its contents and generate much helpful
> > comments
> > for the contributors.
> >
> > The end goal is to let contributors go back and forth against the
> > CI
> > to fix the most obvious mistakes and then let the reviewers take
> > only the final look.
> >
> > Hopefully, it will be a better experience for everyone.
> >
> >
> > [1]
http://frostyx.cz/posts/running-fedora-review-after-copr-build
> > [2]
https://github.com/FrostyX/fedora-review-service
> > [3]
https://pagure.io/FedoraReview/pull-request/463
>
> Jakub, could this service be misbehaving?
>
> See
https://bugzilla.redhat.com/show_bug.cgi?id=2133080#c19
We also saw the same sort of thing yesterday on an update to a review
request:
https://bugzilla.redhat.com/show_bug.cgi?id=2119983
>
> --
> Miro Hrončok
Regards
Frank