Struct LossWatch 

pub struct LossWatch { /* private fields */ }

Watches one device for loss, reported or not.

Implementations

impl LossWatch

pub fn new(monitor: DeviceLossMonitor) -> LossWatch

pub fn unfinished(&mut self, queue: &Queue)

The renderer stopped waiting for what it submitted to queue before it finished (a frame ran out of its completion allowance, or the renderer does not wait at all). Work already watched keeps its earlier start.

pub fn lost(&mut self, device: &Device) -> Option<String>

Why the device is lost, if it is: wgpu reported it, or work found unfinished is still unfinished NOT_PROGRESSING_AFTER later, which this declares. Polls device once, without blocking, when work is pending; costs nothing otherwise.

pub fn stopped(&mut self, queue: Queue, waited: Duration)

The renderer waited waited itself, a bound at least NOT_PROGRESSING_AFTER long, for work on queue that never finished: the GPU stopped. Declares the device lost.