[osg-users] [osgAnimation::TransformHardware]bug BonePalette indices ?

[Julien Valentin]

Okay I found it
It was simply due to stateset sharing among several RigGeometry's sourcegeometries

Given that fact the TransformHardware mod several times the same stateset..and give wrong result

Will make a patch in osgexport in order to handle this case


mp3butcher wrote:
> Hello ol
> Excuse for my incapacity to be more formal but I've hacked a lot osganimation so it's hard to be sure
> 
> But I just reproduce this strange bug, so there seams something wrong with uploaded vertexattrib:
> I add a guard in the skinning VS
> 
> Code:
>  if(matrixIndex < palettsize )
> 
> 
> (palettsize==MAX_MATRIX) ) 
> in computeAcummulatedNormalAndPosition func
> And the render gets badder (so the test gets a false value)
> It means that boneindex is wrong
> (Note that with this guard using uniform or UBO gives the same result so the bug might definitly come from here and not from uniform as i originally state in the subject)
> 
> I add a vertexattribute check to test if the invcondition
> 
> Code:
>  if (indexelement of a vaa>=palettesize && associated weight!=0) print stuff
> 
> 
> on the ram side but  it's never triggered   ?!
> 
> 
> Any idea?
> 
> 
> robertosfield wrote:
> > Hi Julien,
> > 
> > On 23 August 2017 at 00:51, Julien Valentin < ()> wrote:
> > 
> > > Sorry false alarm, I mess up my brain for nothing
> > > Too tired to code
> > 
> > 
> > Does this mean you've got things working?
> > 
> > 
> > Robert.
> > 
> >  
> > 
> >  ------------------
> > Post generated by Mail2Forum
> 


------------------
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=71512#71512