I recently submitted a PR to the epel8-next branch of a dist-git project with zuul-ci enabled. The zuul-ci runs the rpm-scratch-build job against the epel8 koji target instead of epel8-next. This results in a false negative test failure.
epel8-next
rpm-scratch-build
epel8
The standard dist-git-build-pipeline job (which also runs when zuul-ci is enabled and results in duplicate scratch-builds) runs against rawhide which is also wrong.
rawhide
Hi,
Thanks for the report.
The Zuul config does not support epel8-next branch for the moment 1. It seems that the branch matcher 2 is a regexp and the "epel8" matcher matches ("accidentally") the "epel8-next" branch so that Zuul triggered the jobs but again the epel8 target.
@mvadkert what the status of support of epel8-next in TMT ?
epel8-next is basically EPEL for CentOS Stream 8 (?)
That is exactly what it is. When used on runtime, It requires epel8 (you need both). The koji target is called epel8-next and inherits from epel8 but has newer builds of some of the stuff.
Sorry, this slipped through our radar :(
So, seems we are just missing some mapping, to test these against CentOS Stream 8, which we support.