Closed zancas closed 2 years ago
This failure is due to keygen.sh
not running before the test_server
HTTPS tests. The relevant tests need to ensure the existence of the cert-key prior to running.
I've found a few different options for this.. Among them, https://github.com/la10736/rstest and https://github.com/mindsbackyard/galvanic-test
rstest
looks the most minimal of the options I've looked at, so far.
I think it might be better to have our own test.sh
script, rather than needing to instrument all tests, especially as an erroneously un-instrumented test might do something wild (e.g. cause a race condition, which could very rarely and unpredictably cause test failures).
Another (more involved) option is to use our own test harness.
... Or I could just put it inside of create_test_server()
. Duh.
Implementation question: Do we want to rerun keygen every time we run the tests? Or just if there's no key? I'm inclined to say every time. Maybe even use a tempfile.
I say throw it in create_test_server
and throw the file in the test-specific temp-dir. I'm a huge fan of ALL test state being in a test-specific, easily found location.
I see this commit locally: b13c2a22b0ab730308e90437ea53dd902029d6b1