Henri Verbeet [becoming an audio expert now?] wrote:
The default device may for example be a set of USB speakers or an USB microphone, in which case you'd have just an output or just an input respectively.
My testcase has been to hack mmdevdrv.c and replace "default" with "plug:dmix" or "dmix" which is an output only device. Expected test result is to behave like most testbot.winehq.org and skip capture tests. Actual output (prior to your patch) last time I tried was some failures. I.e. the mmdevapi render and capture tests suggest that there's no default device for testbots that cannot capture.
How does your alsa_try_open interact with devices that clang onto the underlying HW resource for several seconds? (PA does this)
Regards, Jörg Höhle
On 22 September 2011 11:40, Joerg-Cyril.Hoehle@t-systems.com wrote:
How does your alsa_try_open interact with devices that clang onto the underlying HW resource for several seconds? (PA does this)
That's patch 2/5, but that aside. It should behave pretty much the same as the existing code for opening devices. It just tests the default device as well instead of blindly adding endpoints for it.