Start at moving tests for git module ⚡️ #462
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #462 +/- ##
==========================================
+ Coverage 60.61% 66.48% +5.86%
==========================================
Files 20 20
Lines 1163 1265 +102
==========================================
+ Hits 705 841 +136
+ Misses 458 424 -34
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
abcclassroom/tests/test_git.py
Outdated
| assert captured_output[0].startswith("Encountered this error") | ||
|
|
||
|
|
||
| # NOTE: Updated to run git status so it works universally |
There was a problem hiding this comment.
@kcranston there is a bigger question here about how much we mock. Right now the big stumbling block is ssh setup for git clone but here i mean git status is easy to just run. Because you run git commit below and it's fine i may update this test to avoid any mocks.
| # and you try to run _call_git? | ||
|
|
||
|
|
||
| def test_clone_repo_pass(monkeypatch, example_student_repo, capsys, tmp_path): |
There was a problem hiding this comment.
this could run IF we could figure out authentication. i really do not understand how to do that part in a way that it works on ci.
|
consider pulling out updates to the code base vs tests... |
This is ready for a review. I still am struggling to raise a filenotfound see #468 and also we need to resolve how we want to handle commit messages: see: #466
#466 #465 are related to this PR and could be addressed here as well potentially
i'm slowly updating docstrings as i write tests.
There is also a general question about some of the git / github functionality. it's easy to test git operations that don't hit github because authentication doesn't seem to be an issue. you can run git status for instance and it just works. But if we use SSH we can't really clone or pull without it benign setup and doing that on ci is tricky.
So one option is to have a suite of tests that only run locally. but even then pushing to a random repo seems a bit tricky too as it does require permissions that a new dev may not have. so i'm not sure how to do this without a good bit of mocking.