Kelsey Gilbert

In particular, having to go through the cpu data path means a couple of extra copies: * 1: getDepthInformation: 1-2 copies * A: (maybe) readback from GPU to cpu *...

There are some concerns about how to upload this to WebGL but I think DEPTH_COMPONENT16 would always work, and would be great for this, allowing both direct sampling of values...

I actually think that getImageData might want to be (effectively) colorspace agnostic, since the goal there is generally "let me work with the backbuffer". toDataURL encodes based on a file...

This is all specified IIRC (de facto where not de jure), but it's a bit of a mish-mash of decisions unfortunately, since we all picked what makes the most sense...

If you could disentangle those, it would help us quickly push through that specification fix!

I think an event is good but it should be orthogonal to accurately presenting any given combo of sdr/hdr. An event could tell you that you may want to upgrade/downgrade,...

Ok, well, that's unfortunate: I think the ergonomics are better this way, as it matches how function feature detection is done, extending the approach there to enums.

WGSL 2024-06-04 Minutes * AB: We don’t have a strong opinion. Two languages did it “sum” and “product” and one language did “add” and “mul”. HLSL uses InterlockedAdd. * **RESOLVED:...

WGSL 2024-05-14 Minutes * Landed [https://github.com/gpuweb/gpuweb/pull/4640](https://github.com/gpuweb/gpuweb/pull/4640) as fix * DN/AB: This relaxation describes current behavior on Windows drivers.

WGSL 2024-05-14 Minutes * (PR is still in-progress) * AB: Regarding naming, naming in WGSL doesn’t use short forms like ‘src’. Was that taken into consideration when designed? * JB:...