fix state-dependent tests #229
Conversation
alanag13
requested review from
antazoey,
kiran-chaudhary,
peterbriggs42 and
timabrmsn
|
CLA Assistant Lite bot All contributors have signed the CLA ✍️ |
|
|
||
| def test_show_when_py42_raises_exception_prints_error_message(runner, cli_state, error): | ||
| cli_state.sdk.cases.file_events.get_all.side_effect = Py42NotFoundError(error) | ||
| def test_show_when_py42_raises_exception_prints_error_message( |
There was a problem hiding this comment.
For CASE_DETAILS, can we make that the actual response for getting a case? And maybe rename it to CASE_DETAILS_RESPONSE_TEXT or something similar
|
can we make the text for the |
| obj=cli_state, | ||
| ) | ||
| cli_state.sdk.cases.file_events.add.assert_called_once_with(1, "1") | ||
|
|
There was a problem hiding this comment.
On line 229, it seems to me like the wrong method is being mocked ... I don't understand what is going on there. get_all does not get called in show, as far as I can tell.
shouldnt we be mocking cases.get() instead?
There was a problem hiding this comment.
the test method at line 229, test_show_prints_expected_data_with_include_file_events_option , mocks file-events get_all, so its fine. When --include-file-events is passed we want all the events listed against the case.
|
Consider it approved on my end, but I found some other oddities and ways that would make the tests less likely to break from py42 changes. |
kiran-chaudhary
left a comment
kiran-chaudhary
left a comment
There was a problem hiding this comment.
Changes LGTM.
Honestly, Looking at the refactor, I am not able to tell what issue has been resolved, some more detail in the description, a link to the nightly build execution, would have helped to understand the scenario.
Having said that, I will try to find the nightly build execution, that shouldn't be hard assuming it should be in last few executions somewhere.
Please ignore this. I started my day with reviews and show the slack conversation later, it did help to understand the investigation part. |
The nightly build was failing due to some tests that only worked when run in a certain order. Going through all tests and running them individually exposed which tests needed adjusting.