USDShader fixes/improvements #6076
Conversation
… to be enum indexes instead of string values.
…i:light:` prefix.
boberfly
force-pushed
the
fixes/USDShader
branch
from
1fb3477
to
c2ca269
Compare
| if allowedTokens and isinstance( plug, Gaffer.IntPlug ) : | ||
| return IECore.IntVectorData( [ i for i in range( len( allowedTokens ) ) ] ) | ||
| else : | ||
| return IECore.StringVectorData( allowedTokens ) if allowedTokens else None |
There was a problem hiding this comment.
Could you give me an example of a shader where this is necessary? I want to refresh my memory and check that this isn't a symptom of a problem elsewhere.
There was a problem hiding this comment.
That one I didn't hit in my testing of MaterialX nodes but the check after it is the one which works for me, so this one might be redundant - the IntPlug check should keep it harmless to leave in though...
There was a problem hiding this comment.
Yep, but what shader can I use to see the problem? I want to check that the problem isn't in the way we're loading the shader rather than in the UI...
There was a problem hiding this comment.
I think I was checking with gltf_surface with the alpha mode plug but there were other shaders I was coming across that needed it, I'll get back to you on a few more tomorrow. I did see from the mtlx schema that they register as integer and not token so I think the loading part was correct (I think token/string is a better approach if that enum list changes though).
Generally describe what this PR will do, and why it is needed
These are the small things I noticed when I added Material X nodes as USDShader nodes, the enumerator presets were a string list in the UI when they should be integer indexes.
ri:light:prefix.I have a build with the Prman USD plugin compiled in, and these extra parameters are on USDLight nodes. This just puts a suffix on the end like the Arnold parameters on Lux lights.
Related issues
Dependencies
Breaking changes
Checklist