SimPosition reporting RelativePosition when avatar is sitting

Its suspose to return a relative position because the avatar is then a child object of the prim it is sitting on. This is how the protocol on the wire works and everyone with libomv clients already does a check to see if the agent is sitting then does a offset and rotation from the parent prim. Fixing this would break any clients that want to actually see where the agent is sitting WRT the prim it is sitting on.

We have code that traverses ParentID hirearchy to find a root prim, I can reproduce the problem, trying to figure out why ParentID doesn't seem to be populated correctly.

With all respect I have to agree to disagree with you Robin. Simposition should report the absolute position of the avatar. RelavitvePosition should report the relative postion relating to SIM as the function name suggests.Otherwise what is the point if they both behave the same???

It is failing cos the code you have is looking for the parent of the parent of the parent....etc. There can be only one parent and not multiple..no? Even if not before this certainly seems to be the case in SIM V 1.40. Sit down on an object on a V 1.40 SIM and you will see what I am referrign to...

Legolas, could you please try LIBOMV-859.patch and see if it fixes problem for you. We're trying to keep compatibility with OpenSim which does this slightly differently.

@Legolas, good point, i retract my statment

In OpenSim, when avatar sits on a child prim, it's parent ID is that of the child prim, so one has to go one more step back to find the root prim. There is some talk about allowing avatars to sit on attachments too.

I've tried to make a universal method, but have unfortunately missed the first step, if there is no hierarchy sigh.

Thx Latif that patch sorted the issue out and I am more than pleased we have a solution that pleases SL as well as OpenSIM. Great work guys. Ty very much