Hi everyone,
I have been thinking if there is a demand I could volunteer to mentor one additional project, for example the Automatic Git Cache Maintenance project idea. I think I will have the time capacity based on my current situation.
That would be great @krisstern ! @rishabhBudhouliya has also agreed to mentor that project idea.
Iâll add your name and his name to the list of possible mentors and will update the status of the project idea of www.jenkins.io.
1 Like
Project idea: implements CD using Jenkins rather than an external concurrent toolâŚ
Although a HUGE topic, it is a very interesting and useful project idea.
I believe it is worth giving this idea some substance (and describe what could be done by a GSoC contributor). Do you have some description/ideas you could throw on a piece of paper and share with us ?
We could then pair together to give it adequate substance so that we can add it to the GSoC list.
WDYT?
there are already some specifications here in a JEP jep/jep/221 at master ¡ jenkinsci/jep ¡ GitHub.
it doesnât look to be so complicated to implement.
But probably the JEP need to be refreshed for some parts or rewrite.
And requires some coordination with Jenkins security and Jenkins Infra team so can be definitely a good exercise for some intern project to have a broader view of what an IT project is.
Basically an alternative to the infra side of JEP-229. Would require setting up a new Jenkins service dedicated to this task with a completely different authorization model pulled from GitHub permissions, changing the RPU processor to push Artifactory bot credentials to that rather than GitHub secrets, and then also securely managing GitHub tokens (perhaps by implementing a GitHub App) to run Release Drafter in CLI mode and edit tags & releases, provide some way of automatically provisioning deployment jobs, and come up with a plan for migrating dozens of existing repositories to the new system. Not at all straightforward (subtle security issues), and certainly not something a student can work on independentlyâwould require close collaboration with someone on the infra team who would need to set aside significant time for this. For plugin developers there would not be much benefit, though we could offer a nicer UX for browsing the occasional deployment failure. To make this not be a pile of one-off hacks I think you would need to overhaul and significantly expand the GitHub Authentication plugin in addition to the jenkins.io-specific infrastructure work.
Thatâs a very interesting amount of valuable informations.
Thanks for that!