10:37 p.m.
On Tue, Feb 21, 2017 at 10:23 PM, Martin Sivak <msivak(a)redhat.com> wrote:
> Finally realizing the potential for harm, we acted quickly to
> re-enable the whitelist mechanism. We made an effort to include
> current authors and maintainers, but we do not know everyone. People
> not included in the whitelist will see the following message when
> submitting patches:
How can the maintainer trigger the CI for a 3rd party patch? We won't
merge anything that didn't pass the tests first so we need a way to
trigger them.
Triggering the patch manually does not work, even if a valid user
trigger the build.
My poor workaround was to upload the next version of a patch.
In vdsm we can also run the tests on travis. After your setup travis
on your fork in github, you can push the change to github to test it
like this:
git review -d patch-number
git push github
Using:
$ git remote -v | grep github
github git@github.com:nirs/vdsm.git (fetch)
github git@github.com:nirs/vdsm.git (push)
Interesting that Travis is not worried about everyone running
CI, no whitelist is needed :-)
--
Martin Sivak
oVirt / SLA
On Tue, Feb 21, 2017 at 6:53 PM, Barak Korren <bkorren(a)redhat.com> wrote:
> TL;DR: If you are a patch author, please ask your project maintainer
> to have you added to the CI whitelist.
>
> The oVirt CI system tries to be a useful and powerful tool for patch
> authors and project maintainers. With the current CI-standards, power
> is placed in the hands of patch authors to make the system do almost
> anything.
>
> We try to be an "open" Open-Source project, so permission is given to
> anyone to open a Gerrit account and submit patches.
>
> The above opens the door to the CI system being maliciously exploited,
> so some countermeasure was needed. The builders of the CI system
> foresaw this and have put in place a white-list mechanism that makes
> the CI system only run jobs for patches that come from listed authors.
>
> In recent years the CI system had been re-engineered with the push
> towards the CI standards, and the whitelist mechanism was rendered
> non-active.
>
> Finally realizing the potential for harm, we acted quickly to
> re-enable the whitelist mechanism. We made an effort to include
> current authors and maintainers, but we do not know everyone. People
> not included in the whitelist will see the following message when
> submitting patches:
>
> To avoid overloading the infrastructure, a whitelist for
> running gerrit triggered jobs has been set in place, if
> you feel like you should be in it, please contact infra at
> ovirt dot org
>
> If you come across this message, please ask the project maintainer to
> send a message to the infra team asking for you to added to the CI
> whitelist. We'd rather not receive direct messages from individual
> contributors, because we do not know everyone and cannot verify
> sources.
>
> We (the infra team) know that the current mechanism can sound
> draconian and be inconvenient. But we had to put something in place
> quickly. Please join the discussion at [1] to improve it.
>
> [1]: https://ovirt-jira.atlassian.net/browse/OVIRT-1154
>
> --
> Barak Korren
> bkorren(a)redhat.com
> RHCE, RHCi, RHV-DevOps Team
> https://ifireball.wordpress.com/
> _______________________________________________
> Devel mailing list
> Devel(a)ovirt.org
> http://lists.ovirt.org/mailman/listinfo/devel
_______________________________________________
Devel mailing list
Devel(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/devel