Will Eastcott
Will Eastcott
Looks like this is triggered from adding a height map to a standard material (for parallax mapping)?
Yes. One option is to create an entity with rigidbody/collision components and create a child entity with a model component and offset the model. Or vice versa.
Can you identify the offending commit specifically?
Sounds reasonable to me @mvaligursky
I like either: material.chunks.vertex.transform material.chunks.fragment.diffuseConst (I hate acronyms like vs/fs) Or: material.chunks.transform material.chunks.diffuseConst (Do we even need to differentiate between vertex and fragment chunks?)
Assigning this one to you @slimbuck, since you're super-keen to refresh PlayCanvas' cubemap handling. 😄
@slimbuck Does the addition of [`reprojectTexture`](https://developer.playcanvas.com/api/pc.html#reprojectTexture) address this? If not, what work would be required to justifiably close this issue, in your opinion?
@mvaligursky I disagree. The idea of `lock` and `unlock` is that they were designed to be called in pairs. Sure, not calling `unlock` is a hack that happens to work,...
Good question. I would probably put something in the docs that you should avoid nesting locks. In reality, it should still work fine if you do. But it does look...
I just hit this issue myself after migrating all of the PlayCanvas Engine's codebase to use class-based JSDoc comments. For example: https://github.com/playcanvas/engine/blob/5251bc18c439d6db4abd757a50f2c6f7f3da50ee/src/math/vec3.js#L621-L627 The doclets are generated with `scope` set to...