It looks like Windows doesn't change the primary buffer length when
changing formats so make the tests fail now.
This means the current ALSA implementation is broken when hardware
acceleration is set to full and the primary buffer format is changed.
Add tests to figure out whether Windows can change the primary buffer
capabilities after setting the buffer format. Based on a patch by Alex
Villacís Lasso.
Added tests for correct Compact behavior.
Added tests for all possible IDirectSound[8] creation modes with good
and bad arguments.
Split dsound.c and ds3d.c into separate files for IDirectSound and
IDirectSound8 interfaces and added code to not execute IDirectSound8
tests when run with a dsound.dll of version 7 or earlier. We should
now be able to run these tests on a Windows system with DX7 or earlier
(untested).
independent.
Set the sound and listener velocity and increase the sound source
velocity so we may get some doppler effect.
Added some traces so one can monitor the listener/sound positions with
WINETEST_DEBUG=3.
fixes a bug where buffer_service() would not detect that everything
had been played.
Check that the actual sound duration was within 10% of the expected
value.
Make BUFFER_LEN and TIME_SLICE independent, make sure TIME_SLICE does
not divide BUFFER_LEN to spice things up a bit.
- the sound format test. Plays a 5 second reference tone to help
debug issues with the tone duration and help detect stutter issues.
Prints explanations describing the sound format interactive
test. Then plays a 1 second test tone for each supported sound
format.
- the 3D sound tests. Performs the volume, panning, moving listener
tests, etc.
Jonathan Ernst