IoT seems to have recently set up a branch that's basically Rawhide, but is versioned 33. We're getting composes under https://kojipkgs.fedoraproject.org/compose/iot/ with compose IDs like Fedora-IoT-33-20200318.0 and the ['payload']['release']['version'] value in composeinfo.json as 33.

This is kind of novel. AFAIR we have never done anything like this before. For the main 'Fedora' compose, Rawhide composes are versioned as 'Rawhide', not as the release number. If you look in https://kojipkgs.fedoraproject.org/compose/rawhide/ , the release IDs are like Fedora-Rawhide-20200318.n.0, and the ['payload']['release']['version'] value in composeinfo.json is Rawhide.

So far, I spotted this because it causes the openQA 'release identification' tests to fail - what they expect to see in the identification files is based on the compose versioning, and this tripped up the logic. I can imagine it might confuse other things too.

It also seems like it'll mean branch points work differently for IoT than they do for regular composes, I guess.

If it's not necessary for things to be this way, perhaps this could be changed so IoT like the main composes has a permanently-existing 'Rawhide' branch?