On Fri, Oct 07, 2011 at 10:35:27AM -0600, Theo de Raadt wrote: > > aucat is similar to X11 in this regard: necessary for a desktop, useless > > on a server. > > Except you are wrong. For instance, my audio box has no display. > > aucat as a daemon is an intragral back-end to a library which more and > more audio programs will be linked to, and those things need to move to > a more powerful API. If aucat is not running, a completely different > audio flow codepath is used. > > Among other things, audio streams have to be mixed by a trusted third > party. One program's library cannot give audio input to another > program to mix it. Classically the over-complex solution has been to > have the kernel mix it. With aucat, the solution is for an ambiant > running daemon to mix it and control everything. It has to be > running, or the old path is used. > > The plan is to gut the direct device code-paths substantially, and > stop trying to perform magic two ways. The direct-device methods will > continue to work, but only as minimally as they did 10 years ago.
I'm totally aware of all of this, I tested sndio when it was not called sndio yet, when ratchov@ (I think) sent the first diff for sdl to misc@. And I read source-changes@. > If some of you people keep insisting on having backwards compatibitity > with the stone age, we'll have stone tools forever. > > I don't know why our mailing lists are always full of people who don't > even understand what they are talking about. Teach yourself before you > try to teach us. Come on, I'm not one of those guys. I never trolled or whined about anything on this list. There was a debate on whether it's good or not to enable aucat by default. I just put forth the idea that maybe you could let the user decide. You don't like the idea? Fine. Best regards, Jona