[ODE] cylinder plane bug

[Bram Stolk]

Geoff Carlton wrote:
> LR wrote:
>>  I don't know if it can help but :
>>
>> - cylinder/plane is ok when encapsulated in a transform geom ...
>>
>>   
> I don't think geom offsets are the culprit here.  In your last example, 
> you can get the same "drop through floor" failure if you just set the 
> rotation of the cyl upside down in reset_state:
>   dMatrix3 mat;
>   dRFromAxisAndAngle(mat,1.0f,0.0f,0.0f,M_PI);
>   dGeomSetRotation(cylgeom,mat);

Your talking about test_cyl, right?

If I add these 3 lines to reset_state() in test_cyl,
test_cyl works just fine for me.
(both single and double precision on linux, svn trunk)

Note that test_cyl does no cyl/plane isects, only cyl and box
vs trimesh.

About cyl/plane isects: I think I'll redo the depth calculation,
which may fix the test_boxstack problem.

  Bram

-- 
Bram Stolk, VR Engineer SARA, Amsterdam.   tel +31 20 592 3000

"Windows is a 32-bit extension to a 16-bit graphical shell for an 8-bit 
 operating system originally coded for a 4-bit microprocessor by a 2-bit 
 company that can't stand 1 bit of competition."