few tests for MongoStorage frontend #542
Conversation
* test with clean mdb * start the test * start test * whoops * maybe too much * Update test_mongo_frontend.py * Update test_mongo_frontend.py * update docs
| @@ -8,6 +8,7 @@ flake8==3.9.2 | |||
| hypothesis==6.17.4 | |||
| immutabledict==2.2.0 | |||
| lz4==3.1.3 | |||
| mongomock==3.23.0 | |||
There was a problem hiding this comment.
This is actually used for the CMT test, but not listed in the requirements. Let's pin it
There was a problem hiding this comment.
Nice test, also very useful to have a working example for testing MongoDB dependent code in the test suite.
Is it actually necessary to have local_sf alongside the mongo_sf? i guess being able to compare the two improves the tests but on the downside, if I break something in the DataDirectory related code during development this test will also fail adding a bit less clear what I broke. Im not sure which approach is better, just wanted to raise it before this is merged and forgotten...
Very good point actually, I've just refactored it a bit, casting it in different test that should fail for different errors without being affected by changes in the DataDirectory |
What is the problem / what does the code in this PR do
Long outstanding issue that code was untested for the MongoFrontend
Can you briefly describe how it works?
Add a test using an empty mongo server from github actions
Result
~95% coverage on the Mongo-storage Frontend/Backend/Saver
Possible issues.
This uses the action secrets which are not available on branches. This may lead to drops in coverage when the test is not run for a PR from a fork that is not AxFoundation (the preferred method of running).A possible solution would be to spin upmongodbdirectlyresolved
Why not use mongomock