I'll try to check this out. I think the problem is the socket-based communication which is really a pain. When a cpu is on powersave mode it appears the sockets are such a low priority at times some messages arrive up to a half-second later... And this is on a lowlatency kernel too...
Once I solidify the pd-l2ork code and release a stable version (and we appear to be darn close to it now), one of the things on my todo list is to do away with socket-based communication in favor of queuing calls via a separate thread. Sure, one will lose gui-less operation but with libpd around, I don't see a point in trying to maintain gui-less implementation (which from what I saw on the list appears to have a fair share of its own problems, please correct me if I am wrong here).
P.S. Don't forget to check out the preset_hub and preset_node (k12 mode uses it so that you can store all states even among multiple instances of same abstraction, which is virtually impossible otherwise). Currently the documentation is not yet integrated (thanks to Jonathan for his work on this) but there are example patches in the complete source inside pd/src folder...
Best wishes,
Ico
-----Original Message----- From: Hans-Christoph Steiner [mailto:hans@at.or.at] Sent: Thursday, October 25, 2012 12:48 PM To: Ivica Ico Bukvic Cc: 'pd-list'; 'IOhannes m zmoelnig' Subject: Re: close all patches on quit sourceforge patch
On 10/25/2012 09:09 AM, Ivica Ico Bukvic wrote:
That said, running core pd-l2ork without all the customized externals will
only
give you a limited picture. E.g. pd-l2ork uses a custom version of cwiid
that
supports wiimote passthrough mode, but that requires installing the custom version of cwiid. If you are using the full installer, this will be done
for you...
To add to this, -k12 flag (or K12 mode) will not work unless you have full thing installed as it relies upon additional files found in the K12 module...
The -k12 mode is nice :)
FYI, what started me on the kick is tracking down GUI slowness bugs.
There
are two that I"m currently working on that affect Pd-vanilla, Pd-extended, and pd-l2ork >= 0.42:
- [bng] stops flashing when data comes in faster than every 50ms
- arrays stop showing updates when data comes in faster than every 100ms
I've almost got a complete fix for the [bng] problem, it now stops
flashing
<10ms, I think I can it working always. The source of the array problem
still
eludes me, here's that thread: http://lists.puredata.info/pipermail/pd-dev/2012-10/018679.html
I attached test patches for both of these problems.
.hc