[pulseaudio-discuss] Working with PA and jack

Patrick Shirkey pshirkey at boosthardware.com
Wed May 27 10:52:32 PDT 2009



Lennart Poettering wrote:
> On Wed, 27.05.09 15:25, Ng Oon-Ee (ngoonee at gmail.com) wrote:
>
>   
>> On Wed, 2009-05-27 at 10:12 +0300, Tanu Kaskinen wrote:
>>     
>>> 2009/5/27 Patrick Shirkey <pshirkey at boosthardware.com>:
>>>       
>>>> So is the only way to communicate with pulse other than brute force kill or
>>>> "pasuspender" to use the dbus protocol?
>>>>         
>>> Probably yes, since it seems that PA's client API doesn't contain
>>> functions for suspending sinks (i.e. release devices). But why would
>>> you want to use something else than D-Bus? This piece of functionality
>>> was developed in cooperation with the Jack developers for just this
>>> use case (or so I believe). Even if there were other ways, Jack would
>>> need to modified to use those ways, so what's wrong with this
>>> particular solution?
>>>
>>>       
>> Just to butt in for a bit on this point, there's a big discussion
>> currently in the jack-devel ML concerning jackdbus. As of this time, it
>> seems to me that its likely dbus support in jack would be optional
>> (plug-in or something of that nature). The upshot of that would be that
>> it would be the decision of packagers whether jack for a particular
>> distro would contain dbus.
>>     
>
> Hmm, unfortunately there are no public archives of that ML (WTF?).
>
> Gah. 
>
>   
>> The issues some have with dbus in the jack-devel ML is basically
>> breakage of legacy as well as some licensing issue with the LGPL jack
>> uses.
>>     
>
> Uh? Compat? That problem is usually much overrated for free desktops.
>
> D-Bus is available under AFL and GPL2. AFL should be fine for LGPL
> uses and GPL2 should be fine for GPL uses. Even companies like Nokia
> who run closed source software on Linux are fine with D-Bus. 
>
>   

After the past couple of days I think Nokia is probably a bad example to 
use for leadership in automated systems design.

There are some very real issues that jack has with being tied 
permanently to dbus and it has been pretty much written off as a compile 
time dependency with the current direction. The aim is to give jack a 
more modular api where dbus can be one of many communication interfaces.





Patrick Shirkey
Boost Hardware Ltd






> Lennart
>
>   



More information about the pulseaudio-discuss mailing list