POV-Ray : Newsgroups : povray.general : Release Candidate? What the ...?! : Re: Release Candidate? What the ...?! Server Time
6 Aug 2024 08:18:18 EDT (-0400)
  Re: Release Candidate? What the ...?!  
From: Tim Nikias
Date: 15 Apr 2002 14:42:41
Message: <3CBB1F38.33385E58@gmx.de>
That was very insight- and helpful.

I do agree that the new POV 3.5 is WAAAAAY better
than 3.1g, and I'm also used to the idea that some bugs
might not be fixable. We have that Microsoft Library
Issue that popped up somewhere using pow(), and some
precision problems. We have that heightfield-normal stuff
and I don't know what else.

But, as requested by your other post (which didn't sound
that comforting, and thus the harsh reply), I created a
"minimum scene" in which I have put my "pet bug" (which
I think is a very nice name for such a terrible thing...) and
made it public in:
povray.beta-test.binaries
entitled as
"Vanish/Ghost/Killer bug"

I traced the problem down to following line:
vnormalize(<0,0,0>)

Using it for translation seems to create that scene-breaking
effect, though POV doesn't issue and problem in dividing
a vector by 0 (I just assume that its done that way:
"vector divided_by vector_length", thats the mathematical
way, right?). Perhaps thats another Microsoft Library thingy?

In that case, I will be content if the docs state something
like "DON'T USE VNORMALIZE WITH <0,0,0>",
and, additionaly, don't forget the "Tim Wenclawiak found it..."

After all, I think it would be nice if every bug-hunter,
which has found an unfixable bug, would be mentioned somewhere,
in some tiny section. We all want to be a part of this, you know...

;-)

--
Tim Nikias
Homepage: http://www.digitaltwilight.de/no_lights/index.html
Email: Tim### [at] gmxde


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.