Vecvec

I've noticed that a validation fix for spv (in trunk now see #5760) was duplicated in this PR, so I've removed that. (I think this was why gfx-rs/naga was requested...

> and tlas packages (which are only in wgpu) being able to have a different max instances from the max instances a tlas was created with (though I can't remember...

@JMS55 I've pushed a commits removes the need for max instance when creating tlas package, but this is a deviation from the spec that this is implementing from. Tell me...

It's mainly a user code thing, we still need the validation (new_with_instances doesn't do this because it can't raise an validation error in wgpu). Personally I just found it annoying...

> I never used the ability to have a different amount of tlas items. there is also still the ability to have a none variant of a tlas instance this...

I have added more detail to some of the docs to make them more like a spec (since this is not in the WGSL spec).

> https://app.codecov.io/gh/gfx-rs/wgpu/pull/6291 suggests there are a few larger methods or branches that just aren't being exercised in the tests, probably worth trying those out. Yes, it seems that `command_encoder_build_acceleration_structures_unsafe_tlas` is...

> I do think we need a little more testing, particularly on the shader side. > We should add execution tests to ensure reasonable behavior for: > - Forgetting to...

> I don't want to overly burden you with testing that code too, we just can currently only do execution tests with the wgpu-level API, so those tests couldn't have...

> Seems we're not minding minAccelerationStructureScratchOffsetAlignment? That's odd, at the top of the file it's forced to 256 (the maximum of any device), so this should work fine... from command/ray_tracing.rs...