First of all thank you for this great example!

It's early in my reading but I'm wondering why not let wgpu choose the underlying device and then write a TryFrom impl or similar for the openxr state. Bear in mind I'm still skimming this repo and have not yet grokked it. Essentially, why go from openxr to wgpu instead of the other way around?

The existing logic does not handle the headset being made available after loss, resulting in the swapchain being used improperly and a resulting crash.

Error logs

Error [GENERAL | xrCreateInstance | OpenXR-Loader] : ApiLayerInterface::LoadApiLayers - failed to find layer XR_APILAYER_LUNARG_core_validation
Error [GENERAL | xrCreateInstance | OpenXR-Loader] : Failed loading layer information
Error [GENERAL | xrCreateInstance | OpenXR-Loader] : xrCreateInstance failed
Error: a requested API layer is not present or could not be loaded
error: process didn't exit successfully: `target\debug\wgpu-openxr-example.exe desktop-with-xr-resolution` (exit code: 1)

Hardware and software

• Valve index
• SteamVR (stable and beta versions tested)
• Windows 11
• Radeon RX 6800 XT
• Driver version 22.5.1

I'm trying to run the example with cargo run -- xr and am getting this error:

PS C:\Users\efsub\source\repos\wgpu-openxr-example> $env:RUST_BACKTRACE=1; cargo run -- xr
    Finished dev [unoptimized + debuginfo] target(s) in 1.32s
     Running `target\debug\wgpu-openxr-example.exe xr`
loaded OpenXR runtime: SteamVR/OpenXR 0.1.0
thread 'main' panicked at 'Unable to load create_swapchain_khr', C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\ash-0.37.0+1.3.209\src\vk\extensions.rs:268:21
stack backtrace:
   0: std::panicking::begin_panic<str>
             at /rustc/897e37553bba8b42751c67658967889d11ecd120\library\std\src\panicking.rs:616
   1: ash::vk::extensions::impl$9::load::create_swapchain_khr
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\ash-0.37.0+1.3.209\src\vk\extensions.rs:268
   2: ash::extensions::khr::swapchain::Swapchain::create_swapchain
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\ash-0.37.0+1.3.209\src\extensions\khr\swapchain.rs:64
   3: wgpu_hal::vulkan::Device::create_swapchain
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\wgpu-hal-0.13.2\src\vulkan\device.rs:559
   4: wgpu_hal::vulkan::instance::impl$5::configure
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\wgpu-hal-0.13.2\src\vulkan\instance.rs:737
   5: wgpu_core::hub::Global<wgpu_core::hub::IdentityManagerFactory>::surface_configure<wgpu_core::hub::IdentityManagerFactory,wgpu_hal::vulkan::Api>
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\wgpu-core-0.13.2\src\device\mod.rs:5089
   6: wgpu::backend::direct::impl$3::surface_configure
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\wgpu-0.13.1\src\backend\direct.rs:1003
   7: wgpu::Surface::configure
             at C:\Users\efsub\.cargo\registry\src\github.com-1ecc6299db9ec823\wgpu-0.13.1\src\lib.rs:3576
   8: wgpu_openxr_example::main
             at .\src\main.rs:136
   9: core::ops::function::FnOnce::call_once<enum2$<core::result::Result<tuple$<>,anyhow::Error> > (*)(),tuple$<> >
             at /rustc/897e37553bba8b42751c67658967889d11ecd120\library\core\src\ops\function.rs:248
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
error: process didn't exit successfully: `target\debug\wgpu-openxr-example.exe xr` (exit code: 101)

I have since installed the LunarG Vulkan SDK, installed the latest drivers for my GPU (Nvidia GeForce GTX 1060) and SteamVR is running before I run cargo run -- xr. Are there any other setup steps I'm forgetting?

With the increasing prevalence of standalone Android headsets, it's important to have this compile and work on the major headsets (Quest 2, Pico 4, Lynx R-1, etc) to ensure that people can see how to build something for the standalone market.

I've bodged the transform from OpenXR space to wgpu space:

let xr_rotation = {
    let o = pose.orientation;
    Quat::from_rotation_x(180.0f32.to_radians()) * glam::quat(o.w, o.z, o.y, o.x)
};
let xr_translation = glam::vec3(-pose.position.x, pose.position.y, -pose.position.z);

This is very much trial and error and not based in any kind of well-understood transformation. Getting to the bottom of this would be ideal.