#1992 allow untagging of stratis modules from rawhide
Closed: Accepted 5 years ago Opened 5 years ago by kevin.

Currently we have a rule to never untag packages/artifacts from rawhide once they have gone out in a compose (ie, never go backwards).

I'd like to ask for an exception to untag 2 modules from rawhide that have already gone out in composes. See https://pagure.io/releng/issue/7740. They were built against rawhide and the normal process tagged them into f30-modular. However, they are meant to be used in f28/f29 which don't themselves have all the build dependencies this module needs. This should just be untagging them from f30-modular and tagging them into candidate tags to allow bodhi updates to be filed for them for f28/f29.

+1s?


I guess +1, but why if it can work in F28 and F29 isn't it also available for F30?

Also, why doesn't the module simply include the missing build-deps on F28 and F29?

@sgallagh, @ignatenkobrain's comment says that the packages are not installable on F30.

Given that the packages are not installable, I think it is safe to untag them from F30 since no users would be able to have them, so I am +1 to @kevin's proposal.

Given that they aren't installable, +1 for the untag

So does bodhi support that yet, tagging the same artifact for f28 and f29?

This seems like a case of https://pagure.io/releng/issue/7662

Bodhi does not support tagging the same artifact for f28 and f29.

Yes, I thought this was planned for f30. So resolving this ticket is not realistic.

Right, bodhi does not support tagging one build into multiple releases, but thats not what we are asking for here. There are TWO builds. One for f29 and one for f28:

stratis-master-20180901144646.7a127764 into f29
stratis-master-20180901144646.337c62b4 into f28

bodhi does support this, they are just two different updates on two different branches. ;)

Ah, I didn't look at the linked issue and misunderstood the context of this ticket -- I thought there was just one build against platform:f30 with requires on platform:[f28, f29] that should go into f28/f29.

Looking at the artifacts and the source modulemd, I see how this works -- clever :)

Okay, +1.

We're at +7/0/0 so I'll mark this as accepted fast track.

Metadata Update from @zbyszek:
- Issue close_status updated to: Accepted
- Issue tagged with: pending announcement

5 years ago

Metadata Update from @zbyszek:
- Issue status updated to: Open (was: Closed)

5 years ago

Metadata Update from @sgallagh:
- Issue untagged with: pending announcement
- Issue close_status updated to: Accepted
- Issue status updated to: Closed (was: Open)

5 years ago

Login to comment on this ticket.

Metadata