Re: [linux-audio-user] is Bristol JACK enabled? - Hallelujah! (kind of)

I am beginning to see that, whatever the problems with Bristol may or may not
be, there are some basic things about how midi is routed in Linux that I need
to understand first. i.e. there are a significant number of different files
in /dev that are midi related, but I don't really know the difference
between all of them. For instance I have not been clear about the
difference between midi, raw midi and the sequencer (it is only from reading
Dave Phillips' post in this thread that I even know that it is the sequencer
we are dealing with in the midi patch bay), nor have I been clear whether,
for instance, /dev/sequencer is an alsa or oss device. Is there a guide
somewhere that explains this kind of stuff?

Another thing: Am I right to understand that whether or not I use jacklaunch
should in itself make no difference to what happens with midi? In other
words will "jacklaunch ./startBristol" and just "./startBristol" do the same
thing? Experiment seems to tell me that, but I want to make sure.

Anyway, this is what happens when I try to launch Bristol in various ways:

$jacklaunch ./startBristol -midi oss
$jacklaunch ./startBristol -midi alsa

Both of the above cause Bristol to launch with no midi-related error messages
and no midi in the patchbay.

$jacklaunch ./startBristol -seq

This is supposed to start Bristol using the alsa sequencer. The application
tries to start up, a window opens but no gui appears, "129:Bristol" shows up
for a second or so in the midi patchbay and then Bristol shuts down.

Here is the output
bash-2.05b$ jacklaunch ./startBristol -seq
        You may want to make bristolengine a suid-root executable
spawning midi thread
Could not reschedule thread to 2
parent going into idle loop
connected to :0.0 (80c53d8)
display is 2752 by 1200 pixels
Window is w 2752, h 1200, d 24, 0 0 0
Using DirectColor display
masks are ff0000 ff0000 ff0000
        flags are 8a000000
midi sequencer
Returning socket 5
Opened listening control socket: 5028
Client ID = 129
Queue ID = 0
Device name did not parse, defaults 128.0
Cannot subscribe port 0 from client 128: Operation not permitted
Error opening midi device /dev/midi, exiting midi thread
INIT: 80c50f0
Initialise the mini link to bristol: 80c9e50
hostname is localhost, bristol
port is 5028
Connected to the bristol control socket: 5
bristolengine already active
80c4f40 80000000 0
parent exiting
return - no data in buffer
cleanupBristol(0)
./startBristol: line 127: 8259 Broken pipe bristol $* -engine
bash-2.05b$

Does that make things any clearer? There are other options I could mess about
with, such as -mididev, but those seemed most obviously connected.

Robert

On March 14, 2005 04:30 am, quoth Dave Phillips:
> Tobias Ulbricht wrote:
> >As far as my experience goes, if Bristol provides an ALSA midi sequencer
> > port, it'll automatically show up in qjackctl. Am I right?
> >If so, jack does not really do MIDI handling/sequencing, I guess.
>
> The MIDI Connections patchbay in QJackCtl is a nice convenience, it's
> not a fundamental aspect of JACK itself. The MIDI patchbay represents
> (IIRC) the status of the ALSA sequencer. If Bristol is not an ALSA
> sequencer client then it will not display in the MIDI patchbay, though
> as a JACK client it will appear in the Audio Connections tab.
>
> As far as I know, at this point JACK has nothing to do with MIDI.
>
> Best,
>
> dp

-- 
Robert Persson
ireneshusband@email-addr-hidden
YahooMess:ireneshusband AIM:shamanicpolice
"No matter how much ye shake yer peg
The last wee drap rins doon yer leg."