"Dan Kegel" dank@kegel.com writes:
Without this patch, winmm/tests/mixer.c appears to hang on one of my machines.
I think the problem was capsA was uninitialized in: for (d=0;d<capsA.cDestinations;d++) {
That's just hiding the bug. If mixerGetDevCapsA failed we shouldn't go on to use the data that it hasn't returned, we should abort the test.