When using Edge post-process detection with WebGPU, it doesn’t work.
I have errors in the console:
babylon.js?1773084184:1 Uncaught TypeError: Failed to execute ‘createTexture’ on ‘GPUDevice’: Failed to read the ‘size’ property from ‘GPUTextureDescriptor’: Failed to read the ‘height’ property from ‘GPUExtent3DDict’: Value is outside the ‘unsigned long’ value range.
cc @UNIT who is the creator of the post-process.
I noticed that the sample from the doc page (linked by @Dad72) generates an error, even in WebGL:
thinEngine.ts:3017 WebGL: INVALID_VALUE: texImage2D: width, height or depth < 0
We don’t have such errors in the original thread (Outline Post-Process), where the PGs use the external mmd package, and it also works in WebGPU: https://playground.babylonjs.com/?webgpu#WGZLGJ#11441. Set useExternalImplementation = false to use the post-process from the post-process library and make it fail.
Maybe there was a problem during the port to the Babylon.js post-process package?
I doubt that UNIT will respond. It’s been 1 and a half years since he last posted here.
No problem, here’s the fix:
Thank you very much Evgeni.