Rotation no longer accepted by space with top of tree sirikata

[danielrh]

I was just testing top of tree kataspace with top of tree sirikata (7fbd3c25a40ccdc010f2723321167bea4826694f) and when you hit create, make a clone of your avatar then rotate it, the rotation is denied by the server and the object spins back to its original location. Translations and scales seem to stick.

the same kataspace code rotates properly with the older sirikata revision

commit 61386211fda6d92ac885c80f7ef61637f15daace
Author: Patrick Horn patrick.horn@gmail.com
Date: Sat Feb 19 00:14:22 2011 -0800

It could well be something to do with a protocol shift, but we should update kataspace to be friendly with sirikata at some point since the old version of sirikata space server had a few outstanding issues that have been fixed in top of tree

[ewencp]

The change in the space server is just that rotational velocity is magnitude-sensitive now, so we may not be getting things normalized in the right places in KataSpace. The point was to have > 2pi radians rotations/second. My guess is that we're computing some unnormalized quaternion when we want a normalized one and then the message sent back from the space server is being respected -- there shouldn't be any 'denial'. The relevant change in Sirikata is here: sirikata/sirikata@ea203b5

[ewencp]

Also, note that this requires pbj changes (or protojs) which was assuming you always wanted your quaternions to get normalized.

[ewencp]

And I forgot to add the relevant link for the pbj changes: sirikata/pbj@acab0c8