On Sep 08 21:18, Jaime Melis wrote: > Hi there, > > One of the pain points I have with using ci.centos.org is that > whenever you want to configure a new repository, you have to wait for > `centos-ci` to accept the invitation. > > This can take a few days, and I usually have to create a bug in > bug.centos.org on behalf of the developers to raise awareness that > we're waiting for this. > > I was wondering if we could improve this workflow. > > I suggest we host a repo under `github.org/centos` with a yaml file > with whitelisted repositories. This way as soon as anyone with > permissions, such as Brian or KB, merges a PR that includes a new > whitelisted repo, it will trigger a job that automatically accepts > that repository invitation on behalf of the `centos-ci` user. I like > this workflow because merges can be done very quickly, and even from a > cell phone ;) It also provides a platform to discuss with the > developers in case there are any questions about the repo, etc, plus > added auditability, etc. > > Note that the old workflow would continue to work, i.e. having the > repo in the whitelist file is not strictly required as invitations can > still be manually handled. But if users actually send a PR it will be > faster for them. Therefore, if we go forward, I suggest we add a note > in this doc about how to accelerate the accepting of invitations by > sending a PR: > https://wiki.centos.org/QaWiki/CI/GithubIntegration > > I put together this script as a PoC: > https://github.com/jmelis/automate-invitations > > What do you guys think? > > cheers, > Jaime > > -- > Jaime Melis > Senior Software Engineer, OpenShift.io > Red Hat > jmelis at redhat.com > _______________________________________________ > Ci-users mailing list > Ci-users at centos.org > https://lists.centos.org/mailman/listinfo/ci-users This workflow looks ok to me. I don't mind accepting invitations this way. Have you counted the number of Github API calls that accept_invitations() makes? I'm curious if it would make 1 or more calls per repo listed in the yaml file, that may be something we can optimize for. --Brian