30 Sep
2014
30 Sep
'14
12:10 a.m.
On Mon, 29 Sep 2014, Fons Adriaensen wrote:
On Mon, Sep 29, 2014 at 08:35:10AM +0200, Giso Grimm
wrote:
Used to be x was launched from the dbus-launch command line, but I think
it is upstart or systemd launching them these days.
I have in the past been successful using dbus-launch to run screen on a
headless box and been able to run pulse from one screen terminal and
jackdbus from another (and other jack/dbus apps in other screen
terminals). Absolutely, there is no need for X to make dbus work. I did
(BTW) find pulse -> jack bridging absolutely useless in this case, except
that it was a great test case to show that dbus was functioning correctly.
There are not a lot of CLI audio apps that look for pulse.
May this be related with dbus support of jack2?
If yes, then using jack1
or recompiling jack without dbus support might help. If dbus support is
needed without having X11 then there was a post on the topic by Len
Ovens ('Headless jackdbus (no X11)' on July 22nd 2013).
Dbus and X11 should not be related. Except that both the 'session'
dbus and X11 are probably launched by the same entity: the *DM I'll be optimistic and assume that jack2 really
doesn't *need*
dbus even if it's capable of using it. Maybe I'm too optimistic.
Even if I've seen too much 'desktop', 'session', 'seat' and
whatever other nonsense creeping into a what should be a basic
system.
Probably not. I don't know if running jackdbus is more or less useful in a
headless case. The major upside I can see is the ability to control the
running jack server from jack_control. Certainly jack_control stop/exit is
much cleaner than killall -9 jackd. If this is worth the added complexity
of having yet another service running, I don't know.
In any case, I would suggest chmod -x whichever executable you are not
using (jackd or jackdbus). Or if you are building jack2 just build the one
you want. (the debian jackd2 package is broken IMO for including both)
--
Len Ovens
www.ovenwerks.net