many sound cards => timeout in winmm:wave tests

Hi,

some native machines on test.winehq time out the winmm:wave tests.
http://test.winehq.org/data/8c00ee0beff1a9a6be08f32d821186c798d15b9d/win7_wi...

Their logs show that they play ~25 times 1 second silence and
a few times 0.5 second silence.  I believe winetest.exe kills tests
that take longer than 30 seconds.  That explains the apparent timeout.

The reason that they play that long is that they are equipped
with several sound cards, each of which being used in turn.

It's good IMHO that all sound cards are used.  However, we should
evaluate whether all combinations of flags that winmm:wave tests
make sense or whether some could be played shorter than 1s.

Or perhaps increase the timeout for winmm:wave?

Regards,
    Jörg Höhle

[André Hentschel]

Am 12.08.2011 15:38, schrieb [e-mail]:
> Hi,
>
> some native machines on test.winehq time out the winmm:wave tests.
> http://test.winehq.org/data/8c00ee0beff1a9a6be08f32d821186c798d15b9d/win7_wi...
>
> Their logs show that they play ~25 times 1 second silence and
> a few times 0.5 second silence.  I believe winetest.exe kills tests
> that take longer than 30 seconds.  That explains the apparent timeout.

i had a quick look at the code and it looks like 120s, still the test lasts really long.
further i think 0.8s is already a better value than 1.0 just by being a fraction.
> It's good IMHO that all sound cards are used.  However, we should
> evaluate whether all combinations of flags that winmm:wave tests
> make sense or whether some could be played shorter than 1s.

we could lower the duration by every cycle of the for loop by some ms until a low limit, that's more likely to happen than:
> Or perhaps increase the timeout for winmm:wave?