ralph opened a new pull-request against the project: `releng` that you are following: `` Patch find_unblocked_orphans for pagure versus pkgdb. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
ralph commented on the pull-request: `Patch find_unblocked_orphans for pagure versus pkgdb.` that you are following: `` This one should be unblocked now. https://pagure.io/pagure/issue/2412 was resolved. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
till commented on the pull-request: `Patch find_unblocked_orphans for pagure versus pkgdb.` that you are following: `` I noticed that the PagureInfo() class does not take the branch into account. AFAIU in the current setup, https://pagure.io/dist-git-requests/tree/master needs to be checked to see which branches are orphaned or not orphaned. For example it might be that only the EPEL branches of a package are orphaned but the Fedora branches are not. Or only Rawhide is orphaned but the previous Fedora releases are not. I am not sure how this is mapped to the new setup. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
till commented on the pull-request: `Patch find_unblocked_orphans for pagure versus pkgdb.` that you are following: `` The ```orphaned_packages()``` function does not take the branch to check into account. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
kellin commented on the pull-request: `Patch find_unblocked_orphans for pagure versus pkgdb.` that you are following: `` @till I do not think the orphan_packages needs a branch name; if I am reading it correctly it seems like a single sanity check to put the package names into a dictionary; not a check of branches.
Essentially - 'does this project exist' - not necessarily a particular branch.
@ralph am I reading it correctly?
Apart from that - it seems OK to me - though I'll be going over this a lot more in depth when I translate this from YUM to DNF in the next few weeks. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
till commented on the pull-request: `Patch find_unblocked_orphans for pagure versus pkgdb.` that you are following: `` @kellin a package is orphaned on a specific branch not in general - at least this was the model in pkgdb, therefore the branch needed to be checked. I am not sure how this changed now. AFAIU the dist-git-requests repo contains or could contain details if only certain branches are orphaned, for example the EPEL branches but not the Fedora branches. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
ralph commented on the pull-request: `Patch find_unblocked_orphans for pagure versus pkgdb.` that you are following: ``
a package is orphaned on a specific branch not in general - at least this was the model in pkgdb, therefore the branch needed to be checked. I am not sure how this changed now. AFAIU the dist-git-requests repo contains or could contain details if only certain branches are orphaned, for example the EPEL branches but not the Fedora branches.
First - real quick, we changed the authoritative repo from `dist-git-requests` over to https://pagure.io/releng/fedora-scm-requests at @ausil's request.
Anyways - we lost the ability to orphan a single branch (like EPEL) with the move to pagure over dist-git. We saw it as a feature that you didn't have to fiddle with multiple people having commit on *all* the branches. I see now that we unintentionally lost the ability to orphan EPEL. I see now what you're talking about in #6957.
The main goal of orphaning per-branch was that we could create a signal EPEL needed a new maintainer (which we don't currently have good tools to do). If that queue didn't get "processed" by packagers, then we would end up needing to retire the branch (which we currently can do with PDC EOLs).
You've been pointing at the override repo as a possible place to store per-branch orphan information. We can proceed with that, but let's think for a bit if there's any other good place to put that. ``
To reply, visit the link below or just reply to this email https://pagure.io/releng/pull-request/6895
rel-eng@lists.fedoraproject.org