> (A quick test showed that taking the square root of a negative number 
within an
 > isosurface does not cause an exception!  That would have been a whole 
other
 > source of frustration; I would have expected the render to crash in 
that case,
 > and therefore would not have suspected that to be the problem if the 
render ran
 > to completion.  Now I have something else to watch out for!)
 >
 >
 >

I've found that a long time ago. I supose that an internal test change 
the negative value to zero, take the absolute value, or there is a trap 
to catch that undefined value.


I suspect that it's the third instance of RE_fn_Blob2...
The first have: (0,y,z, Rpipe)
The second have: (x,0,z, Rpipe)
But the third have: (x,y,z, Rpipe)

It probably should be: (x,y,0, Rpipe)



Alain

Post a reply to this message