`expect`'s count checking error message is inconsistent when multiple `thenReturn` clauses are provided.

[Kyle-sandeman-mrdfood]

Simple example:

def test():
  with expect(xyz, times=1).abc().thenReturn(1, 2):
    func_that_uses_abc_once()

This produces different, seemingly contradictory error messages.

• times=0 or between=(0,0) results in e.g. "Wanted between: [0, 0], actual times: 1"
• times=1 or between=(1,1) results in "Unused stub: abc()"
• times=2 or between=(2,0) results in e.g. "Wanted between: [2, inf], actual times: 1"

After all, how can times=0 say it was executed once, but times=1 says it is unused?

Just to be clear, I am fully aware that thenReturn(1, 2) is the issue 😄 but the error message led me down a rabbit hole around times=1 rather than around thenReturn

Given the len(answers) part in the mockito codebase, maybe a clearer message can be produced if times (and the others) do not match the answer count?
I imagine due to all the chaining that one can do (e.g. .thenReturn(x).thenRaise(y).thenReturn(a, b)) this isn't an easy problem to solve, but if the check_used error message is more clear then one can self-diagnose very quickly.

Maybe something like
Unused stub: .abc(), which also has unused answers?

Once again thank you for the wonderful library!

[kaste]

You're not wrong here. Maybe we can tweak the error message.

[kaste]

Released as 2.0.4

It now says "Only 1 of 2 answers were used..."