[arch-general] PulseAudio in [testing]

[Yaro Kasear]

On Mon, 2010-11-29 at 00:19 +0800, Ng Oon-Ee wrote:
> On Sun, 2010-11-28 at 17:49 +0800, Ray Rashif wrote:
> > On 28 November 2010 11:24, Yaro Kasear <yaro at marupa.net> wrote:
> > > On Sun, 2010-11-28 at 11:17 +0800, Ng Oon-Ee wrote:
> > >> On Sat, 2010-11-27 at 18:21 -0600, Yaro Kasear wrote:
> > >> > I don't see KDE upstream doing that. They have Phonon. What's more, most
> > >> > KDE apps today count on Phonon being there. KDE upstream won't do that
> > >> > without expecting to break KDE.
> > >>
> > >> Admittedly my view on this is skewed, since I follow PA's development
> > >> pretty closely, but one of the devs on PA (Colin Guthrie) has mentioned
> > >> getting various patches he's done for Mandriva's KDE implementation such
> > >> that KDE's mixer and such supports pulseaudio natively. Phonon would
> > >> output directly to pulse in that case, I believe.
> > >>
> > >
> > > The point of which would be what exactly? All due respect, Phonon is
> > > already a sound daemon. To output sound through a sound daemon into
> > > ANOTHER sound daemon, particularly one as poor as Pulse Audio, is
> > > begging for latency and who knows how many other problems.
> > >
> > > And, again, it's redundant and unnecessary since Phonon's already a good
> > > sound daemon on its own merits.
> > 
> > In fact, it handles all my audio pretty well, and even lets JACK take
> > over when needed without my intervention. There's no PulseAudio or a
> > kill command in that equation.
> 
> I'll take both your words on it. Its worth noting that Pulseaudio
> automatically corks when JACK wants a sound-device (jack2 that is, not
> jack1). Running phonon atop pulseaudio wouldn't make sense if every app
> uses phonon. Due to other considerations (for example that all the major
> distros are pushing pulse), this may not be the case in the future.
> > 
> > Anyway, Jan, everything works great, no troubles (with libpulse and
> > without pulseaudio) on KDE. Good job.
> 
> Yes, the lack of complaints (about actual problems) is really
> surprising.
> 

It's probably because the masses of people who already know Pulse Audio
will break their sound aren't bothering to try it. I have a whole IRC
channel filled with people who, if they were to actually test this,
would FLOOD this entire discussion with problems that would make the
Arch devs reconsider this decision.