Catch and verify an expected UserWarning in the test suite

[kurtmckee]

The test suite validates that an SQS FIFO queue with content-deduplication disabled does not receive messages from EventBridge, but doesn't catch and verify the UserWarning that gets raised in the code because of this (recent example in CI). As a result, that UserWarning gets reported by pytest at the end of the test run.

This PR uses pytest.warns() to catch and verify the expected UserWarning.

[bblommers]

Hi @kurtmckee! Thank you for improving our CI logs.

The tests are run twice - once regular, and once in ServerMode.
Moto in ServerMode runs in separate Docker-container, so we can't intercept any user warnings there.

We could handle this with an if-else (if normal: intercept warning & put_events, else: put_events without intercept). I'm not sure if it's worth the effort though, considering it doesn't improve the readability of the test.

If you can think of a different non-invasive way to achieve this, I'm all ears, but the warnings in the CI don't bother me enough to warrant big changes to the test.

[kurtmckee]

Hello @bblommers, and thanks for the quick review!

The benefit to reducing these warnings is to move toward escalating warnings to errors. It is possible to filter out warnings caused by project dependencies, but it's good hygiene to address (or filter out) existing warnings and then -- from a position of zero warnings -- escalate all warnings to errors. This allows CI to catch new Python DeprecationWarnings, or warnings issued by dependencies, early.

It looks like the CI failures are unrelated to this change; were you mentioning the two modes of execution to suggest that this change caused some of the CI failures seen here? Here are the failing tests in CI (Python 3.12, Ubuntu):

• tests/test_core/test_proxy.py::test_http_get_request_can_be_passed_through
• tests/test_core/test_proxy.py::test_http_post_request_can_be_passed_through
• tests/test_core/test_proxy.py::test_https_request_can_be_passed_through

but this PR only touched this test:

• tests/test_events/test_events_integration.py::test_send_to_sqs_fifo_queue

Looking at the Python 3.12 / Ubuntu CI logs, I'm not seeing that this is related, so please let me know if I'm misunderstanding what you're saying!

[bblommers]

@kurtmckee Those tests all reach out to website that was temporarily unavailable - that should be fine when restarting the build. I was talking about these logs:
https://github.com/getmoto/moto/actions/runs/10580339607/job/29315346057

E Failed: DID NOT WARN. No warnings of type (<class 'UserWarning'>,) were emitted.
E Emitted warnings: [].

[kurtmckee]

Aha, I see now. Thanks for pointing that out!

If you're game for it, lemme see if there's something simple that can be done for this scenario (it'll likely be Tuesday morning due to an upcoming holiday), otherwise you can close this PR. 👍

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.49%. Comparing base (5581aa4) to head (8023e34).
Report is 63 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #8041      +/-   ##
==========================================
+ Coverage   94.41%   94.49%   +0.07%     
==========================================
  Files        1134     1150      +16     
  Lines       96832    99025    +2193     
==========================================
+ Hits        91424    93571    +2147     
- Misses       5408     5454      +46     

Flag	Coverage Δ
servertests	28.81% <ø> (-0.19%)	⬇️
unittests	94.46% <ø> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[bblommers]

LGTM - thank you for fixing this @kurtmckee!