#331 moved the tests to a trusted context. Which is great, but we lose visibility of when the test is running and in particular if it fails (as I suspect it did for #330).
How can we handle this better?
Thoughts:
- Report failures back
- Have an action on the PR that runs until the test completes and then sets it's status to the end status?
- Better checks for whether test sites are included correct and fail when not (bonus points - saves us chasing people manually!)
- Something else?
#331 moved the tests to a trusted context. Which is great, but we lose visibility of when the test is running and in particular if it fails (as I suspect it did for #330).
How can we handle this better?
Thoughts: