[LAD] [Jack-Devel] more jack/qjackctl madness

Fernando Lopez-Lezcano nando at ccrma.Stanford.EDU
Mon May 25 03:55:18 UTC 2009


On Sun, 2009-05-17 at 13:11 +0200, Fons Adriaensen wrote:
> On Sun, May 17, 2009 at 12:20:34PM +0200, MarcO'Chapeau wrote:
> 
> > - dbus is not the default compile time config option. Your distro should
> > probably disable it (until it's stable and shiny)
> 
> On this I agree.

(sorry for the delay, I just got back from a couple of weeks of no
internet access). 

Yes, obviously you are not installing from a "reputable" distro. Sigh. I
shall rebuild and disable dbus, I thought at the time I built 1.9.2 it
would be an interesting option and did not realize all the problems it
could/would cause. Argh/sorry.  

I have at the moment still 314 (pi??) more messages marked unread on the
jack list folder in my mail client so most probably this has been hashed
to death already...

Is there a way to disable the dbus functionality if jackmp is compiled
with it? ("jack_control exit"?) I'm far from understanding the whole
thing in detail...

-- Fernando




> > - When dbus is active, it is indeed not backwards compatible with the
> > legacy autostart feature... I agree it might need fixing
> 
> Very very nice. Good luck with it.
> 
> > - The behavior you describe with ports is indeed a bug and needs fixing
> 
> Seems to be qjackctl-0.3.4. Until this is fixed it is completely
> useless for me.
>  
> > And for the record, "fixing a mess" doesn't always need to be "deleting all
> > the features that might create it".
> 
> Depends on what these features are supposed to add.
> 
> The only argument pro using dbus I'v heard so far
> is that it permits run-time discovery of new backends,
> internal clients and their parameters.
> 
> If it were true, it's a very weak argument. We
> do have new backends every day do we ? The last
> one added I know of was more than two years ago.
> 
> But it is bogus. The dbus server uses the existing 
> control interface and it can't provide any info
> that can't be obtained there directly and with
> much less hassle.
> 
> To get backend parameters, you can also just dlopen
> the backend.so and read them directly - there's a
> function for this, used by jackd for its help, and
> also by the control interface.
> Given the name of the backend it takes 10 lines
> of code to do this, which is 3 orders of magnitude
> less than all the dbus stuff required at both ends.
> 
> Dbus assumes there is a local login, without that
> there is no session bus, and things don't work.
> Most of my audio machines are headless, there is
> no local login, but I still expect things to work,
> and that, IMHO, is not unreasonable.
> 
> If it doesn't add any functionality that can be
> provided in simpler ways, and if it doesn't work
> in some perfectly legal use cases, there is no
> reason for having it. 
> 
> Ciao,
> 




More information about the Linux-audio-dev mailing list