Catia missing some ALSA MIDI ports (Seq24)

[Gwyndaf]

Firstly, thanks for the great work on KXstudio. I've one LMDE machine and delighting as each new release brings this into more line with my Ubuntu-based Mint.

I've been using pure ALSA (no Jack) lately, for MIDI hardware, which led me into trying out Seq24. I've done the [manual-alsa-ports] setup to allow external patching, but finding that Catia only shows Seq24 output ports 1-10. 11-16 and Input 1 are all missing. That's more than enough outputs, but the missing input means I can't use Catia to patch Seq24.

aconnect (-i and -o) tells me that ALSA sees all the Seq24 ports (16+1), and Patchage is showing them all too. If I start the ALSA-JACK MIDI bridge in Catia, Jack MIDI also shows all 16+1 Seq24 ports, but still only 10 ALSA MIDI ports. Same behaviour on both LMDE and Ubuntu 12.04 based Mint.

So, I'm wondering whether I'm missing something, or there's a small Catia bug, either in how it talks to ALSA, or displays the information (maybe limit on showing 10 MIDI ports per device)?

It's quite minor, and plenty of workarounds, but just flagging this in case it has wider significance.

Let me know if there's any more information or investigation I can do on this. I tried looking at the Python code, but it's rather above my current skill level.

I'd like to check whether the same behaviour happens with other ALSA software that has >10 MIDI ports, but right now don't know of any, so suggestions very welcome.

Thanks

[Gwyndaf]

Little update on this.

I finally found I can make lots of ports for Qmidiarp with: qmidiarp --alsa --portCount nn (nn=output ports)

This exhibits almost the same behaviour, in that Catia displays no more than 10 ports.

However, in this case it's a little more useful as those ten are the 1 input + 9 output ports.

Which seems to confirm the display limit of ten ports, but odd that it misses the Seq24 input but shows the Qmidiarp one. However, having looked at aconnect, I think it might be because Qmidiarp reports its input as Port 0 while Seq24 reports the input as Port 16. So if Catia shows the first ten ports in numerical order from 0, that would explain this behaviour.

Code: Select all

 aconnect -o
client 14: 'Midi Through' [type=kernel]
    0 'Midi Through Port-0'
client 28: 'ReMOTE SL' [type=kernel]
    0 'ReMOTE SL MIDI 1'
    1 'ReMOTE SL MIDI 2'
    2 'ReMOTE SL MIDI 3'
client 128: 'seq24' [type=user]
   16 'seq24 in        '
client 129: 'qmidiarp' [type=user]
    0 'in              '

[autostatic]

I can confirm this behavior, I've attaced a screenshot.

2014-04-27-220245_930x728_scrot.png (98.02 KiB) Viewed 488 times

[Gwyndaf]

GitHub bug report done. Hope there's enough information, but shout if you need more.

[Gwyndaf]

Great, just received the updates and working a treat now. Big thanks.