winmm/tests: Silence test about play from 0 to 0 not immediately stopped.

Hi,

This intermittent failure seems independent on all the other
intermittent failures so let's silence it.
Experiments with Sleep() were unsuccessful to quiet it.

I'm wondering whether it only happens on vmware.

Another possible explanation that needs more testing is that it might
behave like MCI_CUE: the player starts, does not send a notification
(since it's still running) and is waiting for another, "real", command.
But then, repeatable behaviour would be expected.

Regards,
 Jörg Höhle