Hi Jonathan,
Thanks for the email.
a) Something in the GeoInfo isn't getting cleaned out properly between
filling it with Points and filling it with Polys. From your
description it sounds like the vertex count or point count is somehow
getting messed up and the wrong attribute information is making its
way into the draw code.
All GeoInfos are cleaned out before each drawing by calling
"delete_objects()".
I have the following few lines to set the attributes.
------------------------------------------------------------------------------------------
(1) Attribute* cf = out.writable_attribute(obj, Group_Points, "Cf",
VECTOR4_ATTRIB);
(2) assert(cf);
(3) for (unsigned p = 0; p < num_points; p++)
(4) cf->vector4(p).set(0.8f, 0.8f, 0.8f, 0.5f);
------------------------------------------------------------------------------------------
The strange thing is that with line (3) and (4) commented out, the problem
still shows. That means, by just getting the pointer of the attribute and
without even setting the attributes, there is already a problem.
b) You're not informing the GeoOp hash system that your primitives
have changed by modifying the Group_Primitives hash. If that's not
happening then it's possible the attributes are being left in a
trashed state. Make sure the knob that changes your construction mode
is included in the Group_Primitives hash.
The knob is included in the Group_Primitives hash.
c) There's a bug in the GeoInfo/Primitive OpenGL drawing code that's
not handling the GeoInfo internal change correctly.
Don't know about this. But I created a simple test case modified from
Nuke's
original Sphere example, and it reproduced the problem. Is this the
right place
for me to attach the source code of my test.
Thanks a lot!
-Jean
_______________________________________________
Nuke-dev mailing list
Nuke-dev@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
