1. open a src.fp.o PR that shall fail the Pull Request CI
2. a new pipleine CI job is started, notifies the PR, there is the blue "pending" CI thing with link to the job
3. push force a fix that shall pass the Pull Request CI
4. a new pipleine CI job is started, notifies the PR, there is the blue "pending" CI thing with link to the new job, replaces the old pending link
5. the job from step 2. fails, notifies the PR, there is the red "failed" CI thing with link to the job from step 2.
6. the job from step 4. succeeds, notifies the PR, there is the green "succeeded" CI thing with link to the job from step 4.

Between step 5. and 6. an arbitrary amount of time the "badge" near the PR leads to an outdated failed CI job while the current job is in fact pending. This is confusing but might appear harmless.

Now consider a reversed scenario:

1. open a src.fp.o PR that pass fail the Pull Request CI
2. a new pipleine CI job is started, notifies the PR, there is the blue "pending" CI thing with link to the job
3. push force a fix that shall fail the Pull Request CI
4. a new pipleine CI job is started, notifies the PR, there is the blue "pending" CI thing with link to the new job, replaces the old pending link
5. the job from step 2. succeeds, notifies the PR, there is the green "succeeded" CI thing with link to the job from step 2.
6. the job from step 4. fails, notifies the PR, there is the red "failed" CI thing with link to the job from step 4.

Between step 5. and 6. an arbitrary amount of time the "badge" near the PR leads to an outdated succeeded CI job while the current job is in fact pending and will fail. This can lead to a dangerous situation.