Please try to use more descriptive summaries. The retirement is handled in pkgdb2, there might be already a (closed) ticket about this. It either needs to be discussed at a meeting or maybe pingou can answer it.

I think we orig had this removing all permissions, but then if someone did things out of order they would be unable to update the dead.package file since they no longer had acls.

But I could be mis-remebering...

There are a couple of tickets opened on pkgdb2 for this:
https://github.com/fedora-infra/pkgdb2/issues/86

https://github.com/fedora-infra/pkgdb2/issues/60

The commit message here explains the problem that has been addressed:
https://github.com/fedora-infra/pkgdb2/commit/6f8926fb010e4436229246b778983fc3208ec570

The major problem with removing acls is what happens in case a package is unretired. Especially when a package is unretired within a few days of being retired. (For instance the, "I wasn't reading the warnings about this package being retired but I need it as a dependency of my package so I'll take it" issue or the "I hit the retire button by mistake" issue). Perhaps something can be done to hide the acls in the interface (and package owner email aliases and etc) if the package is retired? Although that leaves people who want to remove their acls for some reason (ie: I was involved in ocaml development but now I'm leaving that community and don't want to be informed of changes whether or not the packages are required or active.) Without a method to do so.

I think I have got the final answer here that the main issue is in pkgdb2 which needs at least single maintainer for the package if the package is not completely getting retired from all the available branches of that package.

This is needed to make sure gitolite talk to pkgdb2 and allows either provenpackager or maintainers to update their own branches. I still think we can add more complexity by checking like if package is retired from any branch but other branch already have (co-)owners then remove all owners from the retiring branch.

Thanks all for providing this information.

Closing this issue.