CI tests failing with panic causing failed tests to not be reported
The check_windows_failures
script is not taking panic
s like this one into consideration. Tests that were supposed to run after the panicked one end up not running, and the script fails to detect and report that/
Proposal
- Check if the
panic:
string is found in any of the test outputs and fail the script with an error message in that case, after accumulating all the panicked test runs, so that we list all of the test runs in one go. - As to the failing test runs:
- If it is expected that they fail in CI:
- Introduce helpers/test package.
- Create a
SkipIfGitLabCI(t *testing.T)
method - Call
t.Skip()
in it if the CI variable is present.
- Otherwise, create a follow-up issue for the failing test and mark it as skipped.
- If it is expected that they fail in CI: