Date: Thu, 23 Feb 2017 12:39:24 +0200
From: Andreas Gustafsson <g...@gson.org>
Message-ID: <22702.48092.78965.364...@guava.gson.org>
| I suppose this is possible, but I think such cases would be rare enough
| that there is no harm in skipping the tests on those.
Sounds reasonable. I will do that (have done that in test sources not
yet in use anywhere...)
| Would the tests even work if there is no audio0?
As best I can see, there are just 2 tests even remotely related to audio
The usr.bin/mixerctl/t_mixerctl test that started all of this - it just
wants the default mixer device (/dev/mixer) - what that requires be existing
in order to work I don't know.
The other is dev/audio/t_pad - that one tests audio0 and pad0 in a rump
kernel - in that environment I assume the test gets to determine that
audio does exist...
The t_pad test is working fine on the i386 b5 tests, and t_mixerctl is not.
Last time the amd64 tests worked (or approx then) t_mixerctl worked there,
and t_pad did not (but that was in mid/late Jan, and may have been while audio
was in a state of flux.)
I should have an updated mixerctl, and t_mixer committed later tonight
(my time ... it is already night here.)
kre
