This page last changed on Jul 23, 2013 by nokk3r1.

Hello,

In newest developer editions KNX DPT 9.001 status command doesn't work. I tested both 21073013 and 18072013 versions and same story. This affects my temperature setpoint sliders. Tested on win7 and RPi. On 17062013 it still works.

Thanks for reporting, and testing the development releases. Will take a look at it.

Posted by juha at Jul 24, 2013 07:54

Figured out that bug could be in range sensor and not necessarily knx or dpt related.

Posted by nokk3r1 at Jul 25, 2013 07:06

Jussi,

Tracking down issues caused by the KNX patches here: Signed 8Bit integers for KNX DPT 6.xxx.

Trying to narrow it down.

Posted by juha at Aug 05, 2013 10:09

Jussi,

I need to review a set of KNX related patches that were applied on 2013-06-25 so after the 2013-06-17 that you reported was working correctly but before the 2013-07-18 you reported broken.

I will need to re-apply the patches one-by-one to find which one got broken functionality. There were 5 patches, unfortunately quite tangled so it was easier to merge them all in at once.

I've gone back to a baseline revision R8778 that should show identical behavior to the 2013-06-25 you had working. Will start adding patches on top of that. If you can find time to test these patch binaries individually and report back issues with DPT 9.xxx sensors, that would help a great deal.

Base binary: OpenRemote-Controller-2.1.0_SNAPSHOT-R8778.zip

Posted by juha at Aug 06, 2013 10:14

I try to find some time to test it this evening or tomorrow..

Posted by nokk3r1 at Aug 06, 2013 10:27

Thanks.

Posted by juha at Aug 06, 2013 16:40

The five patches that I'm tracking down are:

ORCJAVA-291
ORCJAVA-294
ORCJAVA-301
ORCJAVA-361
ORCJAVA-362

The 2013-06-17 snapshot already included ORCJAVA-361 and ORCJAVA-362 but those had to be reapplied due to changes introduced in ORCJAVA-294.

There's a rollback binary built on the base R8778 binary (posted above) that adds ORCJAVA-291 (RGB) to the patch set (so should be ORCJAVA-361-362-291): OpenRemote-Controller-2.1.0_SNAPSHOT-R8778-ORCJAVA-291.zip

If the new rollback base binary R8778 above is not showing the regression, the binary with ORCJAVA-291 applied on top of it would be next one to try.

Posted by juha at Aug 06, 2013 16:49

Both r8778 and orcjava291 works as they should.

And if I remember correctly it was just that "STATUS" command which didn't work so that sliders or labels couldn't get any values, but by moving sliders the actual setpoint was still altered so the "TEMP" command did work. And as you know, they both use same DPT. Don't know if that helps tracking down the problem.

Posted by nokk3r1 at Aug 06, 2013 18:44

Thanks for testing it.

I will create one more binary with ORCJAVA-301 included in it if possible. If it still shows no regressions, then I can isolate the DPT issue to a single patch and drill down on it.

– Juha

Posted by juha at Aug 06, 2013 18:54

Happy to help. Just downloaded 18072013 version again and have to report that it also deploys panel way longer than older builds: every knx command takes about 3 seconds to load and since it deploys them through several times(three I think) it takes a while. And that is same story on both win7 and raspian. Though that doesn,t matter that much since I have dedicated PI for ORC and I can wait 5mins..

And just tested that behavior is as described above: the temp command works in that dpt

Posted by nokk3r1 at Aug 06, 2013 19:07

Yes, this is the side-effect of ORCJAVA-294 discussed here: Signed 8Bit integers for KNX DPT 6.xxx.

The workaround to the startup delay seems trivial but I'm trying to get rid of the DPT9.xxx issues first so I don't have to chase unrelated reports.

Posted by juha at Aug 06, 2013 19:12

Sorry, haven't noticed that thread, since I have done mostly outdoor works last week as weather has been great.

Posted by nokk3r1 at Aug 06, 2013 19:20

The last binary to test would be with the patch ORCJAVA-301 applied, it is available here: OpenRemote-Controller-2.1.0_SNAPSHOT-R8778-ORCJAVA-291-301.zip.

It seems unlikely this last patch would cause the breakage but better test to make sure. If there are no regressions, then the issues can be isolated to a single patch.

Thanks again for testing.

– Juha

Posted by juha at Aug 06, 2013 23:55

That version also works as it should.

Posted by nokk3r1 at Aug 07, 2013 06:31
Document generated by Confluence on Jun 05, 2016 09:40