Open source 3D rendering solution adhering to the RenderMan standard

Bug

slxargs library bug

Posted by: mosari

Hi Guys

 I try to interrogate a compiled shader using slxargs library but my program halt with access violation error.

 I use win32 platform and vc 8 to compile my program.

 To find the problem I reduce my program to something like this:

  char* myShader = "smoeshader.slx";

  Slo_SetShader(myShader);

  Slo_EndShader();

  Call stack said me the access violation happened in Slo_SetShader() function. I

Should CqSubdivision2::limitPoint use catmull-clark curve subdivision for border vertex?

Posted by: sopvop

Hi,

Since prman now interpolates facavarying boundary, I needed to make correct UVs for texture artists. While looking for info on catmull-clark subdiv I stumbled on Aqsis documentation. I see what it uses 4/6*v + 1/6*e1 + 1/6*e2 for border point calculation, prman seem to use 4/3*v +1/8*e1 +1/8*e1 (catmull clark curve). Is it intended behavior in aqsis or a bug?

More SDS detail issues

lancelet's picture

Posted by: lancelet

Hi Everyone,

I've experienced some issues recently with subdivision surfaces (SDS) exported from Blender via Mosaic.  The problems appear to be connected with SDS level-of-detail calculations within Aqsis.  I'm using the current SVN Aqsis (will switch over to GIT as soon as that becomes official), compiled using gcc 4.0.1 under OS X 10.5.6, and the development branch of Mosaic.

This is best explained with some example images (below):

Missing include in SVN 2728

WHiTeRaBBiT's picture

Posted by: WHiTeRaBBiT

Hey guys,

Just thought I should let someone know I had issues compiling SVN 2728 under Linux. The compiler complains that "memcpy" is not a member of "std" in "aqsis/texturing/texlib/iffoutputfile.cpp". I resolved the issue by adding "#include <cstring>" to the top of the file. Anyways just thought I'd share in case anybody else runs into this ;-)

Cheers,

Eric Back (WHiTeRaBBiT)

Possible sidedness bug - affects Mosaic

lancelet's picture

Posted by: lancelet

Hi Everyone,

I've come across a possible bug in the Aqsis handing of single-sided primitives.  The problem shows up in the following RIB: