kmix PulseAudio support

Christian Esken esken at kde.org
Thu Jan 7 00:43:04 GMT 2010


Am Donnerstag, 7. Januar 2010 00:59:17 schrieb Helio Chissini de Castro:
> On Wednesday 06 January 2010 17:09:17 Christian Esken wrote:
> > > > sprint multimedia will be a ground zero point to new mixer interface
> > > > and
> >
> > What exactly do you mean with "new"?!?
> >
> > You guys get me quite confused in this mail thread. Please do not say "it
> >  is bad" or "lets change things" without indicating what you mean.
> 
> Moin Christian

Hi Helio.

Foreword: Please everybody understand: I am really happy that KMix gains so 
much attention at the moment. I hope you will not be offended by some of my 
rants. But keep in mind: new things are cool, but we have a terrible amount of 
bugs in the new things.

> I think is time to separate backends and frontend

In terms of different processes? I thought about it during the design for 
KDE4, and gave it up for the following reasons:

1) You will have to pass horribly much through DBUS
2) The delays will likely suck
3) It will have a very negative impact on battery usage for mobile devices.
 At least for Backends w/o poll(). And actually ALSA is not 100% poll(), as we 
use a timeout for the poll().

Probably I am wrong here. But please take some time to think about it.

> Our interface is too old.
Old is not a quality criteria.

My greatest concern is: Keep it small, keep it fast, keep it simple.

> Our backends in other way are pure beauty.
> So, in the sprint multimedia i will try to raise the bar and make the
>  current visual part go away, make backends controlable by some callback
>  system ( of course dbus )

The Backends have always been  reachable by DBUS interfaces - so what is your 
plan? I would agree that it is time to review the current interfaces.


>  then free kmix to have ideas for new interfaces,
>  and even mutable interfaces for different devices.
> 
> Is a bold idea, but it's time to do now.

Yes. But fixing bugs would be quite bold too. Just changing GUI options won't 
fix too many of them, and introduce other bugs. Unless we get the bug level to 
a normal number (e.g. 20-30), we shouldn't add too many new features.
--- rant starts here ---
We got  bugreports of crashes due to new backends (OSS4), because of Plasma 
hype (introduces screen flickering with OSD), global shortcut issues (the 
whole time since KDE4) and a whole shit of other bugs due to non-mature KDE4 
base and "cool" features I wouldn't have added.

So if there are new features, I do require new components in the bugzilla, 
with ownership of the commiter/supporter. 
--- rant ends here ---

  Christian







More information about the kde-core-devel mailing list