On 2010-11-23 12:42, Lennart Poettering wrote: > On Tue, 23.11.10 08:08, David Henningsson (david.henningsson@xxxxxxxxxxxxx) wrote: > >>> The second patch includes a minimal udev rule (executed synchronously, >>> enabled on all systems, regardless of systemd is used or not) and two >>> systemd service files (executed asynchronously at boot/shutdown, only >>> enabled if built with systemd support). >> >> Lennart, >> >> Could you clarify how this affects/changes behaviour for distros not >> running systemd? > > As I wrote above it won't even be installed if systemd is not around. Ok. >> Also, I personally don't like the idea of saving the mixer state at >> shutdown. (Although I know Ubuntu has that as well.) Too many times >> I have cranked up the volume for some reason, then turned the >> computer off, only to find the login sound at maximum volume. > > Then disable it. And since you don't run systemd you won't even have to > do that explicitly. Ok; so the question is whether we should even have such a rule, and if so, should it be enabled by default? Do other people than you and me have an opinion here? >>> Given that most big distributions are moving to adopt systemd sooner or >>> later >> >> That assumption is not necessarily true at this point. AFAIK, it >> remains to be seen. That said, I don't mind a systemd patch into >> alsa-utils, as long as it doesn't break or bloat anything for >> non-systemd distros. > > Well, this thing is far less "bloaty" It is not bloating at all (at least not by my standard) if it's not even installed, don't worry. > then what Ubuntu is currently > doing. i.e. Ubuntu duplicates the mixer init database in an init script > /sbin/alsa-utils which is moved there for weird reasons. I can only > encourage you to drop that "bloat" and rely on the mixer init database > that comes with alsa, like everybody else. > > (Sorry, the word "bloat" just sets me off) > > But anyway, if you don't have systemd, you won't even see these > files. Nothing changes for you. > > (Also note that Ubuntu is currently asynchronously executing said init > script from an udev rule. That means that by the time a device added > event reaches PA the mixer has not been configured properly yet, which > can confused PA quite a bit. Hence please make sure to drop the > execution of said init script from the udev rule, either by simply > adopting the code this patch includes, or by making your current rule > synchronous. Thanks.) Thanks for the tip, reducing upstream delta is always a good thing. IIRC some of it is inherited from Debian, so it wouldn't hurt for me to find out a little more about this at that level. -- David Henningsson, Canonical Ltd. http://launchpad.net/~diwic _______________________________________________ Alsa-devel mailing list Alsa-devel@xxxxxxxxxxxxxxxx http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
