Struct WgpuContext 

pub struct WgpuContext {

    pub device: Device,
    pub queue: Queue,
    pub surface: Option<Surface<'static>>,
    pub config: Option<SurfaceConfiguration<Vec<TextureFormat>>>,
    pub headless_texture: Option<Texture>,
    pub target_width: u32,
    pub target_height: u32,
    pub depth_format: TextureFormat,
    pub surface_view_format: TextureFormat,
    pub msaa_samples: u32,
    pub anisotropy_clamp: u16,
    pub render_path: RenderPath,
    pub pipeline_settings_version: u64,
}

Core wgpu context holding GPU handles and render target state.

This struct owns the fundamental wgpu resources needed for rendering:

• device: GPU device for resource creation
• queue: Command submission queue
• surface (optional): Window surface for on-screen presentation
• headless_texture (optional): Offscreen texture for headless rendering

The render target dimensions and format are always available via [target_width], [target_height], and [surface_view_format], regardless of the active mode.

Fields

device: Device

The wgpu device for GPU operations.

queue: Queue

The command queue for submitting work.

surface: Option<Surface<'static>>

Window surface for on-screen presentation (None in headless mode).

config: Option<SurfaceConfiguration<Vec<TextureFormat>>>

Surface configuration (None in headless mode).

headless_texture: Option<Texture>

Offscreen render target for headless mode (None in windowed mode).

Created with RENDER_ATTACHMENT | COPY_SRC usage to support both rendering and GPU-to-CPU readback.

target_width: u32

Render target width in pixels.

target_height: u32

Render target height in pixels.

depth_format: TextureFormat

Depth buffer format.

surface_view_format: TextureFormat

The view format used for render target views (always sRGB-qualified).

msaa_samples: u32

anisotropy_clamp: u16

render_path: RenderPath

The active render path. Stored for runtime branching in the frame graph.

pipeline_settings_version: u64

Version counter for pipeline-affecting settings (HDR, MSAA, RenderPath). Incremented when these settings change, used to invalidate L1 pipeline cache.

Implementations

impl WgpuContext

pub async fn new<W>( window: W, init_config: &RendererInitConfig, settings: &RendererSettings, width: u32, height: u32, ) -> Result<WgpuContext, Error>

where W: HasWindowHandle + HasDisplayHandle + Send + Sync + 'static,

pub async fn new_headless( init_config: &RendererInitConfig, settings: &RendererSettings, width: u32, height: u32, target_format: Option<TextureFormat>, ) -> Result<WgpuContext, Error>

Creates a headless (offscreen) GPU context without a window surface.

This constructor skips surface creation and instead allocates an offscreen wgpu::Texture as the render target. The texture is created with RENDER_ATTACHMENT | COPY_SRC usage, enabling both rendering and GPU-to-CPU readback.

Arguments

• target_format — Desired render target format. Pass None to use the default Rgba8UnormSrgb, which provides broad compatibility and straightforward CPU readback. Use Rgba16Float for HDR workflows or Rgba32Float for maximum precision.

pub fn resize(&mut self, width: u32, height: u32)

pub fn set_vsync(&mut self, vsync: bool)

Dynamically reconfigure the surface present mode for VSync toggling.

This is a no-op in headless mode where no surface exists.

pub fn create_depth_texture( device: &Device, width: u32, height: u32, format: TextureFormat, ) -> TextureView

pub fn size(&self) -> (u32, u32)

Returns the current render target dimensions.

pub fn is_headless(&self) -> bool

Returns true if the context is running in headless (offscreen) mode.