Pavel Raiskup
Pavel Raiskup
Thank you, @greyli
Sorry @greyli, I forgot ... I've added you to the PyPI project now. Would you mind adding a git tag and finishing the 0.9.1 release?
OK, I'm closing this. The host's RPM stack is "too new" - it is expected that the database is not readable by the older RPM stack "in chroot". The bootstrap...
Hmm, sorry, I still can not reproduce this. When mock creates those directories, it creates them so `mock` group has access there (`mock` group ownership). Is this still an issue?
I see, no problem. Thank you for the question anyway!
Createrepo fix [was merged](https://github.com/rpm-software-management/createrepo_c/pull/325). - we should have a new option per/project, something like "Only latest NEVRAs in repository metadata" - we need to wait for the createrepo_c release
> What is the benefit of not doing this by default? Some reasons from the top of my head: - createrepo_c is slower when used like that (eats more memory)...
- Some users might want to take care explicitly about the possible duplicates.
JFTR, https://github.com/pulp/pulp_rpm/issues/2271 would solve this for us for free once we are moved to PULP.
Please take a look at #2420, I promissed some time ago. This is the first step; second step should be to announce this and turn it off even for existing...