On Fri, 2009-10-02 at 09:57 +0200, Lennart Poettering wrote: > On Fri, 02.10.09 03:58, Matthew W. S. Bell (matthew at bells23.org.uk) wrote: > The logic behind the actions for sound card devices is a bit > weird. Have you seen the comments at the top of > /lib/udev/rules.d/78-sound-card.rules? Yes. > If those rules are not installed then things will not work. Which is > one of the reasons we require a brand new udev in PA. Last time I > checked Debian didn't have udev 143. Does it have that now? If it > doesn't then it is not exactly surprising if things don't work for > you. Yep, using udev 146, passing the configure check. > Also, for some weird reason Marco modifies the default udev rules > heavily. Which is really something he shouldn't do. I shall add that to the already long list. > > To test this, I ran 'udevadm trigger', which resulted in both cards > > being detected. However, I now cannot reproduce the original problem. > > This only triggers "change" events, which is what PA looks for. Try > triggering "add" events. Er, yes, doing this works, because they are change events. The point is that when module-udev-detect starts, it use the udev_enumerate interface to get udev devices, and this is not a monitor interface. Therefore, there is NO action associated with the device, so nothing gets triggered in process_device(). I think. Matthew
