Re: [LAD] [Jack-Devel] more jack/qjackctl madness : some comments

PrevNext  Index
DateSun, 17 May 2009 20:18:55 +0200
From Stéphane Letz <[hidden] at grame dot fr>
ToPaul Davis <[hidden] at linuxaudiosystems dot com>, Nedko Arnaudov <[hidden] at arnaudov dot name>
CcLinux Audio Developers <[hidden] at lists dot linuxaudio dot org>, JACK Developers <[hidden] at jackaudio dot org>
In-Reply-ToPaul Davis Re: [LAD] [Jack-Devel] more jack/qjackctl madness : some comments
Le 17 mai 09 à 20:10, Paul Davis a écrit :

> On Sun, May 17, 2009 at 2:04 PM, Stéphane Letz <[hidden]> wrote:
>>
>> The point is that when compiled in D-Bus mode, libjack behaves  
>> differently
>> regarding the way it start the server: it does not use the fork 
>> +exec mode
>> anymore but call the D-Bus service to start the server. This  
>> "simple" change
>> is the source of all the problems we then see.
>
> so if i understand correctly, there is effectively a layer of
> indirection. rather than a request arriving via D-Bus leading to a
> normal fork-exec, it leads to D-Bus service request, which presumably
> (somewhere, sometime) leads to a fork-exec of the server? is this
> correct?

"Show me the code" :

http://trac.jackaudio.org/browser/jack2/trunk/jackmp/posix/ 
JackPosixServerLaunch.cpp

It starts the D-Bus jackaudio service that actually starts "jackdbus"  
process with it's own logic (server settings save/restore and so on...)

Hoppefully I described it right (Nedko again?)

Stephane
PrevNext  Index

1242584376.10970_0.ltw:2,a <76E3B169-A4BF-4701-9697-5AED11A7B03B at grame dot fr>