[PATCH 3/3] winmm: Audio drivers unconditionally invoke DriverCallback

[Joerg-Cyril.Hoehle at t-systems.com]

Hi,
[this is the old 6/6 unmodified]

The testcase below doesn't belong to the attached patch #3, it's just
To show another test I conducted on testbot to analyse what sets of
invalid callback flags don't prevent open from succeeding.

In summary, particular situations are detected by open,
e.g. non-null non-window handle, most others are ignored (as in this test).


Sadly, winmm/tests/wave.c is an unreadable mess.  I believe it
would help if it would remember one piece of valid WAVEFORMAT
where open succeeded and subsequently modify it, instead of using
fixed parameters like 22050/16/2 which are not guaranteed to
work everywhere.