Move sources from DistGit to GitLab and confirm the feasibility of the build system and related CI elements running from the location of the newly relocated source.
With Pagure being deprecated for a while now and with people moving away to GitLab as their default location for housing their projects, we need to ensure that GitLab is capable of housing the package sources effectively and the workflow changes that might come into the picture as a result of them being relocated from Pagure to GitLab.
It would not be as streamlined of a request as the Pagure to GitLab importer one as we would want to ensure that we not only relocate our package sources but we recommend changes in the workflow to extensive use of GitLab-only features and avoid vendor lock-in.
We would want to integrate Zuul CI to check for the validity of the source files, Packit to automate the building and testing of packages and many more changes to the packaging workflow to help improve the packager's quality of life. We would also want to account for version monitoring and deeper integration with the APIs on Bodhi and/or Anitya.
What are your nice or really nice to have wishes?
Will this initiative impact CentOS, Fedora? All users? All contributors? A group of contributors (which)?
Packagers.
For Fedora, to begin with, but if possible (and/or required) CentOS ones too.
Is this initiative under a time constraint? Should it start or end before a certain date?
Not that I can think of.
This initiative goes a lot more beyond just zuul and CI. It touches to how we want to manage admin access to gitlab projects, how we plan on people being able to self-serve their own needs (adding new collaborators, adding people to EPEL only or Fedora only branches...) as well as the integration built on other tools (bugzilla and anity/the-new-hotness being two)
The initiative request is valid though, however your comments have given me an idea. Do you think it would be possible to tackle this project in sets or phases? Can we break it up to, for example, - Build system setup - Packager Workflow - CI Enablement - Project Support & setup (EPEL, ELN, etc)
Just an initial thought, whats your read on it?
The initiative request is valid though, however your comments have given me an idea. Do you think it would be possible to tackle this project in sets or phases? Can we break it up to, for example, - Build system setup - Packager Workflow - CI Enablement - Project Support & setup (EPEL, ELN, etc) Just an initial thought, whats your read on it?
The investigation and planning for sure, the actual implementation/roll out, not sure.
Note: I have my thoughts on how we could make gitlab work, happy to share them, but I also don't want to share them too early to avoid introducing a bias to the people looking into it.
Metadata Update from @t0xic0der: - Issue assigned to t0xic0der - Issue tagged with: Accepted, In Progress
Following the 2024 Git Forge Evaluation https://communityblog.fedoraproject.org/2024-git-forge-evaluation/, the investigation has now taken the form of Dist Git Comparison https://fedora-arc.readthedocs.io/en/latest/dist-git-comparison/.
The investigation has concluded on the Fedora Council side of things.
Here are some publications to cement the completion.
Now, we gotta get to the actual stuff.
Log in to comment on this ticket.