krohmerNV

Fixes for the 3-way comparison in tests_to_html.py

2

Tried to do a comparison between MDL 1.6, 1.7, and 1.8 and found two issues: - the computed file paths are not correct when passing different `--inputdir1`, `--inputdir2`, and `--inputdir3`...

Test Suite results for pbrlib/bsdf/wedge_conductor are not consistent

1

The wedge_conductor material shown in the image has an anisotropy of 0.0. Yet OSL and GLSL produce something anisotrop. https://github.com/AcademySoftwareFoundation/MaterialX/blob/c34a1066bb194d36174454b3b763a6d68bc8480d/resources/Materials/TestSuite/pbrlib/bsdf/wedge_conductor.mtlx#L6 

CompareAnisotropy seems to have conflicting normal and anistropy direction

5

This is about one Material in the CompareAnisotropy model, the `grooved-anisotropy` material. The material looks like this: ``` { "extensions": { "KHR_materials_anisotropy": { "anisotropyStrength": 1, "anisotropyRotation": 1.6 } }, "name":...

WIP: add MDL Compiler and Render

1

Draft to add the MDL compiler and a renderer to the repo to simplify development and testing. Note, everything can be discussed. I just started with the approach I find...

In the MDL code generator we currently can't handle layering constructs like this: `layer(mix(bsdf1, bsdf2, weight), bsdf3)` so I propose a workaround that should not change the result but uses...

The PR contains tests we separated from #2125. The goal of this PR is to find a general approach to test backend specific features.