Closed PythonCoderAS closed 4 years ago
I don't believe this is an issue in general. In #1403 you modified the behavior to change what used to be a single request into multiple. That change led to rate-limiting behavior kicking in, where it hadn't before.
What's more, since re-recording cassettes is fairly rare, it's easy enough to just update any patching when that is done on a case-by-case basis.
I'm saying that it's a good future-proofing technique.
Right, I hear you. I'm just saying that those tests didn't need to be patched until you updated the code that it was testing. There's no use in patching a test before it needs it.
I'm just marking it is an issue so that it can be used to track patching efforts.
There are no tests that call sleep
but do not patch it — after all, if there were, those tests would fail. And tests that do not call sleep
should not patch it.
Re-recording cassettes does not cause sleep
behavior to change. Changing the underlying PRAW behavior is what causes that. Please close the issue.
I see, it's a fundamental misunderstanding on my part. Sorry for wasting your time.
No worries! I see where you're coming from.
An effort should be made to patch every integration test. There are a lot of tests that only work because they have not been rebuilt. While that makes it work, if any significant revision leads to the rebuilding of tests, then all of the new tests will fail.
Originally posted by @PythonCoderAS in https://github.com/praw-dev/praw/pull/1403