Fix the section name in wait_for_oidc_server function

[npalaska]

We missed this change earlier, our wait_for_oidc_server function wasn't working at all. We use [openid] as a section name in our config file but the wait_for_oidc_server function was still using [openid-connect] which results in openid not configured error and we were silently ignoring it.

[dbutenhof]

While it seems that the test case you removed was flawed before, and breaks differently now that you added the missing return... do we have test coverage of the failing OIDC connect case?

[webbnh]

do we have test coverage of the failing OIDC connect case?

I believe that Nikhil fixed the unit test so that it will now correctly detect the presence of the successful call (to the mock), and, likewise, it will detect the absence of such a call; however, apparently we don't have a test which exercises the case in which the wait fails.

It would also appear that we have no test coverage for wait_for_oidc_server().

So, @npalaska, are you up for adding some unit tests?

BTW, @npalaska, here are two more spots you might want to touch up: endpoint_configure.py and load_keycloak.sh

[npalaska]

however, apparently we don't have a test which exercises the case in which the wait fails.

We do have a test called test_main_wait_for_oidc_server_exc that test for a generic exception raised from the function but yeah we don't have a specific test case for OpenIDClient.NotConfigured exception which is the only exception we catch and return 1.

It would also appear that we have no test coverage for wait_for_oidc_server().

do we have test coverage of the failing OIDC connect case?

I will add some test coverage for wait_for_oidc_server.

[webbnh]

Nice work, Nikhil! (💯 coverage on shell.py!!)

Remaining questions:

• Do we want more or different handling of the timeout in shell.py?
• Shall we tighten up the endpoints stuff?

Or, should we just declare victory?

[dbutenhof]

Unfortunately, Mr Jenkins has expressed their displeasure:

Traceback (most recent call last):
File "/var/tmp/jenkins/tox/py39/lib/python3.9/site-packages/pbench/test/unit/server/auth/test_auth.py", line 456, in test_token_introspect_aud
assert str(exc.value.cause) == "Invalid audience", f"{exc.value.cause}"
AssertionError: Audience doesn't match
assert "Audience doesn't match" == 'Invalid audience'

• Invalid audience

• Audience doesn't match

[dbutenhof]

Unfortunately, Mr Jenkins has expressed their displeasure:

Traceback (most recent call last):
File "/var/tmp/jenkins/tox/py39/lib/python3.9/site-packages/pbench/test/unit/server/auth/test_auth.py", line 456, in test_token_introspect_aud
assert str(exc.value.cause) == "Invalid audience", f"{exc.value.cause}"
AssertionError: Audience doesn't match
assert "Audience doesn't match" == 'Invalid audience'

• Invalid audience

• Audience doesn't match

For the record, though, this isn't "you" ... Siddardh's and Riya's builds failed the same way. We have a new latent bug or broken package dependency, apparently ... 😦

[webbnh]

It's a thing of beauty!