pfNode::isect() traversal path problem

New Message Reply Date view Thread view Subject view Author view

From: Jason Daly (jdaly++at++ist.ucf.edu)
Date: 01/25/2002 16:51:44


Hi, I've posted on this problem before, but we've recently gained a little
more knowledge about it.

The original problem showed up only on IRIX machines (not PC's). After a
call to pfNode::isect(), when we queried the traversal path and it was deeper
than 32 nodes, only the first 32 would be valid pointers in the pfPath
structure, all others would be NULL. If we call getNum() on the path, it
returns more than 32.

This problem wasn't as evident on the Linux version of Performer. However,
this week we've run into the same problem. On the PC, it more often occurs
when the pfPath is deeper than 64 nodes. In this case the pfPath then
contains the first 64 nodes in the path, with the rest NULL. Occasionally,
we've seen it happen with 32 nodes as well. In both cases, a call to getNum()
on the path indicates more than 64 or 32 nodes in the path.

This is the third time I've posted on this problem. Can anyone on the
Performer team comment on this? It would be nice to know that it's being
investigated :-) I'd be happy to provide any more information you need.

Thanks,

--"J"

"I'm a castaway stranded in a desolate land,
 I can see the footprints in the virtual sand."
        --Neil Peart


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Fri Jan 25 2002 - 16:50:43 PST

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.