[Qmidiarp-devel] Re : Some feedback on QMidiArp 0.4.1

[Frank K. <goe...@ya...>]

Hi Louigi, hi Brendan,

>________________________________
>De : Louigi Verona <lou...@gm...>
>À : qmi...@li...
>Envoyé le : Jeudi 2 Juin 2011 14h45
>Objet : [Qmidiarp-devel] Some feedback on QMidiArp 0.4.1
>
>
>Hey everybody!
>
>Just got my hands on QMidiArp 0.4.1, using it on Ubuntu 10.04.
>This is a fantastic piece of software, however, lack of stability and a "random stop" bug make it virtually unusable in its current state. I would like to report both problems.
>
>1. Random crashes.
>QMidiArp crashes often, in my experience always as a reaction to an action, not by itself. Actions, however, are different - it may be moving a slider or creating a new module. Crashes happen fairly often, rarely have I closed QMidiArp myself, usually it just crashes within 3-5 minutes of work.
>
>2. Random stops.
>QMidiArp, after playing for 15-30 seconds would suddenly stop. The play button is pressed, so in order to restart you have to click on Play to change it to stopped state and then press play again. After 15-30 seconds QMidiArp stops again. In case this info could be useful, the first time I started QMidiArp, it did not stop and I actually played with it for several minutes until it crashed due to me moving a slider in one of the modules, but all subsequent uses of the program featured random stopping.
>
>
>I am available to test the app. It is a very promising tool and I would be happy to help out any way I can. I do dream to actually start using QMidiArp on stage as soon as possible.


Thanks for your reports, this is very important, alas I cannot reproduce crashes on my system. I cannot remember a single one actually, so Louigi does this happen in a situation of sync to jack transport or even when running with the internal clock? Does it crash even with stopped transport ?
Brendan, do you get crashes, too or only the random stop thing?

on 2) I've pushed a fix for a transport sync problem that was present when only arp modules were in the setup, which was definitely a bug that came in since 0.3.9

I've had a discussion with Jeremy "Autostatic" on the IRC, and he found that a 1000Hz system timer (kernel option) and lower jack latencies (around 256 points, 2 periods) worked quite well for jack transport sync.

So could you guys please checkout from git repo and tell me if this solves some issues?

I'd like to know more about the crashes, maybe we'll have to go through gdb etc...

Cheers and thanks again for reporting so far

Frank