CI: Shared runners are disabled for our repo; makes CI jobs for people forking our repo stuck
- The shared runner is disabled for our repo, because it's too slow.
- If people are forking our repo, the CI / CD settings are copied as well.
- Therefore, people are stuck with, by default, a repo configured to run CI jobs, but without a runner which runs the jobs.
- Effectively, this means that GitLab will parse the CI config, create jobs and then waits for a runner to pick up the jobs.
- Because there is none, after some time GitLab will run into a timeout, kill the jobs, merging will be blocked.
The related MR tries to improve the situation to at least document it.
The proper way would be to talk to upstream to try to find / develop a more sane solution.