Learn more about these different git repos.
Other Git URLs
For epel10, the built SRPMs and RPMs have el10_0 dist tag. But fedpkg assumes the dist tag is el10 (without _0).
So the original SRPM has the wrong disttag (which is fine, mock will rebuild it and generate a correct SRPM) but if I do koji wait-repo --build=$(fedpkg verrel) epel10.0-build it will never finish as that package will never be available
koji wait-repo --build=$(fedpkg verrel) epel10.0-build
I was looking into implementing this and fedpkg. Between Carl and I the only idea we can come up with to solve this is by pulling data from Bodhi, but nothing in __init__.py where the functions for this live are talk to Bodhi - all of that happens in cli.py.
__init__.py
Having __init__.py directly instantiate a BodhiClient looks like it will be pretty messy and inconsistent with existing code so I've not PRd anything for it.
I'm curious to see how this gets implemented.
This will also likely break the "check if NVR was already built in Koji" functionality of fedpkg build (the thing that can be disabled via --skip-nvr-check), as it will always say it wasn't built.
fedpkg build
--skip-nvr-check
I think I see a way to fix this. ELN uses Commands._get_build_macros to set a few properties. That fetches macros set on the build tag of the target matching the branch name. We can use the same approach to get our distcore value. I plan to put up an implementation of this but it will work best as changes on top of #560, i.e. after #560 is merged.
Commands._get_build_macros
@dherrera fixed this in #562, which has been built and published as fedpkg-1.45-5.
Log in to comment on this ticket.