Struct Buffer 

pub struct Buffer { /* private fields */ }

A GPU buffer that can act on itself — built via BufferBuilder, never constructed directly. Carries its own device/queue access internally, so writing to it doesn’t need a &wgpu::Queue threaded in from the caller.

Opaque by design: there’s no way to reach the underlying wgpu::Buffer from outside this crate. Binding one into a bind group goes through BindGroupBuilder, which accepts &Buffer directly.

Implementations

impl Buffer

pub fn write(&self, data: &[u8])

Overwrites this buffer’s contents with data, starting at offset 0.

pub fn write_at(&self, offset: u64, data: &[u8])

Writes data into this buffer at a byte offset — for updating one element of a DynamicBuffer without touching the others. Prefer DynamicBuffer::write_element, which computes the offset for you from the buffer’s own stride.

pub fn size(&self) -> u64

Size in bytes.

pub fn read(&self) -> impl SpawnableFuture<Vec<u8>>

Copies this buffer’s current contents back to the CPU. The copy itself is submitted eagerly, right away — do not call mid-frame; call after presenting or outside of frame encoding. Only the wait for the GPU to finish mapping it is deferred into the returned future.

This doesn’t run itself — drive it with AsyncEventWriter::spawn to get the result delivered as an event, or BackgroundTasks::spawn_async directly if you’d rather hold onto a TaskHandle and poll it yourself.

pub fn read_as<T: Pod + Send + 'static>(&self) -> impl SpawnableFuture<Vec<T>>

Same as read but the resolved bytes are cast to T.