bpo-17013: Extend Mock.called to allow waiting for calls

[Kentzo]

This implementation of bpo-17013 is alternative to #16094

Changes are based on my work for asynctest. Specifically on _AwaitEvent that was left out when related code was ported to CPython.

Key features:

• Gives meaning to the existingMock.called property, otherwise not much useful
• Does not require end users to do anything: change is automatically available in every Mock
• Utilizes existing semantics of python conditionals (both asyncio and threading)

Accepting this change will allow me to port _AwaitEvent therefore giving identical semantics to both wait-for-calls and wait-for-awaits.

I will provide necessary typing annotations for typeshed.

Considerations:

1. This approach changes type of the Mock.called property from bool to private bool-like _CallEvent. However, the only practical problem I could think of is if someone was checking type of Mock.called (e.g. via isinstance). That does not sound like a plausible problem: after all the property itself is seldom used with exception of conditional expressions where _CallEvent.__bool__ and _CallEvent.__eq__ is sufficient.
2. It probably makes sense to provide convenience methods like wait_for_call, but I would like to hear the opinion of the reviewers.

CC: @vstinner@tirkarthi@mariocj89

https://bugs.python.org/issue17013

[Kentzo]

Probably one method that the object should definitely have is wait_for_call(self, call, /, skip=0, timeout=None).

[mariocj89]

Thanks @Kentzo for putting this through. From conversations with @voidspace, @lisroach, @tirkarthi and in the issue tracker it seems that people wanted to go with having a different mock just for this use case, rather than trying to fold it into the existing one. Similar to what was done with AsyncMock.

Happy to re-evaluate it, but I guess that should be a conversation in the issue tracker.

[cjw296]

Closing this PR for now, as it looks like @mariocj89 will be putting in another PR to take over from this.

[Kentzo]

@cjw296 Hmm, where was that discussion?

[cjw296]

https://bugs.python.org/issue17013#msg352211
#16094

Looks like @mariocj89 's PR actually predates this one, I don't have a strong opinion on either, so happy to re-open this PR if you, @mariocj89, @tirkarthi, @voidspace and @lisroach want to discuss which PR to take forward.

[mariocj89]

Hi @Kentzo, this was initially discussed in https://bugs.python.org/issue17013#msg352211 and offlne in the core dev sprints with @voidspace, @lisroach and @tirkarthi when we put forward the other PR. @voidspace preferred to go down the path of having specialized mocks rather than having a mock have all the functionality. Similar to the asyncmock.

[Kentzo]

At some point I was asked to move the discussion to bugs.python. For the benefit of the future developers I believe it makes sense to point out the benefits of the new "class approach" there as well.

[mariocj89]

At some point I was asked to move the discussion to bugs.python. For the benefit of the future developers I believe it makes sense to point out the benefits of the new "class approach" there as well.

Absolutely, that's why I suggested that. The issue lays on the availability of committer time though. It might be a busy time right now.

[Kentzo]

Added implementation of the wait_for_call convenience method that can be used like this:

mock.method_1.called.wait_for_call(call(1, a=1), timeout=0.01)

[cjw296]

@Kentzo - it's unclear where you've added this?

[Kentzo]

I guess PR is not updated because it's closed. The corresponding branch in my repo got updated.

[Kentzo]

Any chance for this PR to get reviewed?

[cjw296]

Doesn't look like I can re-open this PR, so please submit a new one if you want anything reviewed. Please comment here with a link to the new one.

@mariocj89 - where did you get to with your approach?

[mariocj89]

There is an open PR that was reviewed but never re-reviewed after proposed changes here: #16094

[Kentzo]

@cjw296#20759

[Kentzo]

Just noticed that I put wrong link above 🤦🏻‍♂️, fixed the comment.