From https://pagure.io/fedora-ci/general/issue/24
the (of a Pull Request) head is created by an async task leading to potential race-condition between the CI system and pagure depending on the load on the pagure instance.
My concern is that the CI is an async job triggered by opening the PR. And so is the head update. The CI will usually take longer to set up before it comes to fetching the head, however this sounds very fragile. Rebellions are built on hope, not CIs.
@bookwar could we check if the two are in sync before proceeding with CI somehow?
Login to comment on this ticket.