Would it make sense to un/block it even if it is blocked somewhere in inheritance chain? It would behave in more similar way as package blocking and it will remain blocked even if it will be from some reason unblocked in inherited tag. Now it is not possible due to test on line 1597.

Btw, we don't have this call available via CLI, which could be worthy.

Would it make sense to un/block it even if it is blocked somewhere in inheritance chain? It would behave in more similar way as package blocking and it will remain blocked even if it will be from some reason unblocked in inherited tag. Now it is not possible due to test on line 1597.

Well, unblock happens in a different function.

The grplist functions are set up very similarly to the pkglist functions, minus the policy check. Can you give examples of the behavior being different?

You can get around the check on line 1597 by using force=True.

It differs in following example where all package operations are valid, nut unblock group fails:

Tag structure:

tag_a
 └─ tag_b

koji add-pkg tag_b package
koji block-pkg tag_b package
koji unblock-pkg tag_a package

koji add-group tag_b group
koji block-group tag_b group
koji call groupListUnblock tag_a group -> results in 'group group is NOT blocked in tag tag_a'

Oh I see. Thanks for the clarification

I think this inconsistency is somewhat independent of the issue at hand. I've filed #895 to track it.

In the meantime, I think this PR does resolve the bug at hand. Unless there are objections, I'll plan to merge it this week.

Commit 4d0fed0 fixes this pull-request