Improve CI tests for specific cases #4791
Description
Whenever you encounter a bug, it's evidence that CI didn't catch the problem, and very likely it's not being tested at all.
That's a good chance to increase test coverage.
| OS | type of runner |
|---|---|
| Windows | ephemeral jit |
| Windows | ephemeral non-jit |
| Windows | standard non-ephemeral |
| Ubuntu | ephemeral jit |
| Ubuntu | ephemeral non-jit |
| Ubuntu | standard non-ephemeral |
and what should the test be?
- print "Hello World"
- check that the username is the configured username
Which test would be failing currently?
I believe "Windows" would fail the username check, and always appear to be "System". There's an open PR now to address this. Is it difficult to include ephemeral windows runners in current test suites? It should not greatly increase costs, since they are ephemeral.