AssertNotCalled() panics because mutex is nil

[gkdr]

hello,

thank you for your continued work on this library.

i have an issue that for some reason only came up today, but i believe this is a similar issue to #1206. i have one test which only checks that a function was not called using AssertNotCalled(), which now fails because the mutex is nil and testify panics.

i instantiate the mock using new, is that wrong?

i am willing to prepare a PR for you, but i am not sure what the fix is. in #1207, a check is added:

	if m.mutex == nil {
		m.mutex = &sync.Mutex{}
	}

does this need to be added to all of AssertNotCalled(), AssertCalled(), AssertNumberOfCalls()? is that all or are there any other places?

[PrafulFP]

I think a fix is also needed for IsMethodCallable and MethodCalled.

[Kryvchun]

The same about AssertNumberOfCalls

[psmarcin]

Using yourmock.Test method before yourmock.AssertNumberOfCalls populates mutex. It's just temporary workaround because this change looks like breaking change for me.

[yurishkuro]

Having a similar issue - upgrading to 1.7.4 causes some tests to panic.

Just curious, what was the motivation for changing mutex to a pointer in #1182 ? The PR did not provide any reasoning, and considering that the mocks are usually initialized as plain structs, having a pointer is asking for trouble.

[peczenyj]

I think a fix is also needed for IsMethodCallable and MethodCalled.

Indeed, I have a m.Called() without On on my programs -- the called was not expected, should fail, not panic on nil reference

[boyan-soubachov]

Fixed in latest version (1.7.5), please confirm and let me know if all is good :)

[PrafulFP]

Working fine now. Thanks 👍🏻

[gkdr]

reverting the change to a pointer fixed the issue. thank you!