[ODE] Fwd: Trimesh troubles.

[Kris]

Hmmmm

Double check if memory allocated for vertices and indices is not being
deallocated somehow... i've had similar error in dCollideBTL. Then
exception has been thrown: cannot read from address 0x45323443 (more less
:P) - what matters is that address wasn't near 0x00000000, so it was
pointer to my deallocated memory :]


Have no other idea. There must be bug in your code somewhere, because
trimeshes works pretty fine...

On Thu, 17 Mar 2005, Derek Arndt wrote:
> >>> 			uint indexCount = meshes[k].data.triangles.size();
> >
> > Number of indices is 3 times larger than triangles :], so you have too
> > few
> > indices in your vector.
>
> Actually, that name is deceiving.  Triangles is an STL vector storying
> type unsigned int (so for each triangle I push back three indices), so
> the size should be correct.
>
> > Remember that when you create trimeshdata pointers to vertices and
> > indices
> > are being copied, veritces and indices vector are NOT being copied, so
> > when youcreate next trimesh you will overwrite vertices and indices of
> > previous trimesh :P
>
> Yes.  This is from a class that is created for every collider, so the
> data doesn't change unless the model changes (and I call this
> function), which I currently don't do.
>
>
> Anyways, things still don't work quite right with trimeshes and I have
> friends telling me they've not gotten them to work either (on OS X),
> I'm losing hope and time :(.
>
> Derek
>
> /*
> http://www.dereksucks.com    _game development is just my forte_
>  >81
> */
>

Pozdr __DEVELOPER_TRINITY_CREEZ_H__
-- 
===========================================
                 GG: 1271378
-------------[ PROJECT TRINITY ]-----------
  [ http://trinity.glt.pl ]   [ Trinity ]
        [ The post-nuclear cRPG ]
===========================================