mav on renderworld()

Blitz3D Forums/Blitz3D Programming/mav on renderworld()

b32(Posted 2006) [#1]
In my texpaint program, I get an error on RenderWorld()
So I wrote a small test version using different methods, but it keeps giving the same error.
Is there something I'm doing wrong ?
I would appreciate it if someone could try to run this code and tell me if the bug occurs on their machine. It takes a minute for the bug to appear.
Has someone experienced the same problem ? I've read a few posts on this error, but in that cases people made too many triangles on a surface.

When I made a version that handles single surface meshes, there is no bug. Same goes for the original texpaint by birdie. So I think the problem could have something to do with using PickedSurface().


skidracer(Posted 2006) [#2]
Works fine for me.


b32(Posted 2006) [#3]
Thanks skidraces, I appreciate it.
The trouble with this bug is that is takes really a while for it to appear.
I've traced the bug using a counter and speeded the example up using "Flip False".
When counter reaches 1995, the bug appears.


skidracer(Posted 2006) [#4]
If there is a chance a texture is missing you probably want to test if tex%=0 before you go and lock / setbuffer the texture, if that doesn't help comment out the call to oval, if that fixes it it sounds very much like a driver bug...

If still no, then can you remove the millisecs() dependency and use an incrementing variable and repost so I am running the same test with the same numbers as yourself.


b32(Posted 2006) [#5]
I think it is strange that I've had this similair problem in two versions. I tested it on another PC, and there it gives an error, too.
Removing the oval did solve the problem! Also when I removed the two 'SetBuffer' commands: no error.

The bug appears after a fixed number of frames.
PC1:2330 PC2:3200
It doesn't matter where I enter the 'loop'. Maybe it is a problem with releasing something ?

All other bugs with RenderWorld() involved some kind of memory leak, for instance too much triangles on a surface.

GetSurfaceBrush creates a new brush, but the texture applied to it, is it a copy too ? I checked if the handle of this returned texture is the same as the original handle, and it is the same.


b32(Posted 2006) [#6]
I still think it is a bug. Sigh .. I will find a workaround.


bytecode77(Posted 2006) [#7]
it is not crashing to me, but sometimes blitz' renderworld crashes on to many triangles!


b32(Posted 2006) [#8]
Hmm. Thanks for testing it, oh child of the devil. I think Skidracer is right then, about it being a driver problem. I have a NVidia board on both my pc's. Thanks again Skid and sorry I had my doubts. I can avoid the error by creating a new texture and draw on that.