Dzmitry Malyshau

Alternatively, build something on https://github.com/krolli/vk-parse/tree/master/vk-parse

We should do it after #1532 since there is going to be infrastructure to forcing the baking of more expressions.

This looks like a bug in validation layers? If they catch a bug in the application, they are supposed to spew out a readable message.

This is now a blocker for [VSand](https://github.com/Ealrann/VSand)

> Dota doesn't discover some of the features we have I just noticed that it does indeed print the following lines when ran on gfx-portability: > Vulkan physical device (0):...

@caiiiycuk you are confusing two different things here. `wgpu_core::id::DeviceId` is a unique identity of a logical device in wgpu's internal storages. But the `uint32_t deviceID;` is not that. Instead, it's...

Thank you for tiling this bug! Something doesn't match up to me. Documentation for [vkCmdSetLineWidth](https://www.khronos.org/registry/vulkan/specs/1.2-extensions/man/html/vkCmdSetLineWidth.html) doesn't list the requirement (of `VK_DYNAMIC_STATE_LINE_WIDTH` being set on the current pipeline) in valid usage....

Thank you, that sounds about right, and I was missing these details. So what do the examples do? Bind a non-line pipeline with this dynamic state, setting the width, and...

Good points! Let's move the line width state out of the `PolygonMode::Line` then in gfx-rs.

This is basically https://github.com/gpuweb/gpuweb/issues/1867. It's a super strange rule that I'm not happy about, and we want to encode it into actual syntax in WGSL. cc @jimblandy