winmm/tests: Properly recover from the WAVERR_STILLPLAYING case.

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

Hi,

waveOutClose is incorrect since 1.3.25 because it closes a playing device.

That's not what I'm patching now, rather than the attempt to
gracefully handle the situation where STILLPLAYING is detected during
the tests.  AFAICT the only way to ensure waveOutClose succeeds is
to precede it with waveOutReset.

I've also reduced some playing times in the hope that some native machines
with 3 cards or more won't enter time out.

Regards,
 Jörg Höhle
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-winmm-tests-Properly-recover-from-the-WAVERR_STILLPL.patch
Type: application/octet-stream
Size: 4728 bytes
Desc: 0001-winmm-tests-Properly-recover-from-the-WAVERR_STILLPL.patch
URL: <http://www.winehq.org/pipermail/wine-patches/attachments/20120119/97838053/attachment.obj>