RPM Support

this skip list will lead to the situation when modular rpms will be synced and become ursine.
Ursine modular rpm can break client's system.

This would make sense only in case rpms are skipped along with modulemd.

+1 to Ina's point.

That skip functionality should be smart and skip related RPMs as well.
If there is no real demand for that feature, I wouldn't add it just for the sake of consistency.

1. Do you guys do the "smart" behavior for errata ?
2. I could see this feature useful if some repo or cdn has broken module md while good packages. I would agree it being a lower priority.
3. Where does this word "ursine" come from ....

1. Do you guys do the "smart" behavior for errata ?

No, we don't.
RPMs and an Erratum where they are referred can be both present in a repo but they also can live separately:
- only RPMs are synced - it's fine,
- only Erratum is synced - it's fine (it can be on purpose, so user selectively adds RPMs to a repo later). Only Errata and no RPMs is quite useless but harmless,
- both RPMs and an Erratum are synced - it's fine.

For modules it's critical to do it smart way:
- only a module is synced and no RPMs - it's bad, it's a broken module, tools like dnf will complain.
- only modular RPMs are synced - it's bad, even worse than the previous case, because it can mess up dependency resolution (modular rpms will be used when non-modular should be) .
- a module and all its RPMs are sync - it's the only good option.

2. I could see this feature useful if some repo or cdn has broken module md while good packages. I would agree it being a lower priority.

I understand this potential use case, however see my comment above, it's not worth risking, in my opinion. If RPMs are good they won't be re-downloaded, just checked during a sync.

3. Where does this word "ursine" come from ....

Modularity folks use it a lot. I guess it describes how bad/damaging and clumsy modular RPMs can be when they don't have their module in a repo ;)