Don McCurdy

> What's your reasoning for making the BatchedMesh keep a handle to the original geometry? I was thinking of the `setMatrixAt( i, matrix )` function (similar to InstancedMesh), and needing...

Thanks @gkjohnson! I'm happy to take their word for it and skip vertex baking, or at most consider it a short-term solution. I think you're suggesting an `N * 16`...

> > I think material parameters per geometry support is needed. > > I strongly disagree with this. If people need different per-object material properties there are a variety of...

We really need to avoid `three-*` imports unless we are actually going to publish those NPM packages. If we're not sure, then it would be best to just use a...

I've been thinking of these aliases as shorthands to make our recommended workflows (namely, using ES Modules) more convenient for external users. But I consider `examples/jsm/libs/*` to be internal dependencies,...

That is a workflow I'd prefer to discourage. By paving a path toward what is ultimately a fragile configuration, I think we are doing new users a disservice. three.js should...

I'd missed #23348 and the removal of unit tests on `examples/jsm/*`, but for posterity here are tests on SceneUtils.sortInstancedMesh: ```javascript import { sortInstancedMesh } from '../../../../examples/jsm/utils/SceneUtils.js'; import { BufferAttribute }...

Agreed! I'll add documentation for `sortInstancedMesh` if the PR is accepted.

Probably these are the cases to consider: | aoMap.texCoord | uv2 provided | result | |-------|-----|--------| | uv2 | ✔ | ok | | uv2 | | invalid | |...

I'm certainly in favor of the change — my only reservation is about which of these three changes can be done independently and which should be batched into a single...