Update to 0.13!

The change log above contains most of the details about what has changed about WGPU and therefore the tutorial. I will make a special mention about how to use map_async() as that has changed. Previously map_async returned a promise that you had to await before you could access a buffers contents. It now expects a 'static callback that takes the Result of the mapping attempt as a parameter. This means that if we want to save a buffers context to an image instead of doing the following:


{
    let buffer_slice = output_buffer.slice(..);

    let mapping = buffer_slice.map_async(wgpu::MapMode::Read);
    device.poll(wgpu::Maintain::Wait);
    mapping.await.unwrap();

    let data = buffer_slice.get_mapped_range();

    use image::{ImageBuffer, Rgba};
    let buffer =
        ImageBuffer::<Rgba<u8>, _>::from_raw(texture_size, texture_size, data).unwrap();
    buffer.save("image.png").unwrap();
}
output_buffer.unmap();

We would do the following:

{
    let buffer_slice = output_buffer.slice(..);

    let (tx, rx) = futures_intrusive::channel::shared::oneshot_channel();
    buffer_slice.map_async(wgpu::MapMode::Read, move |result| {
        tx.send(result).unwrap();
    });
    device.poll(wgpu::Maintain::Wait);
    rx.receive().await.unwrap().unwrap();

    let data = buffer_slice.get_mapped_range();

    use image::{ImageBuffer, Rgba};
    let buffer =
        ImageBuffer::<Rgba<u8>, _>::from_raw(texture_size, texture_size, data).unwrap();
    buffer.save("image.png").unwrap();

}
output_buffer.unmap();

You can explore the reasoning for the change to map_async in this PR.

Another thing to note is that presentation modes need to match what the surface supports. You can get a list of supported modes from Surface::get_surface_modes().

As always, let me know if I missed anything in the migration process. You can open an issue/submit a PR on the Learn Wgpu repo!

2.2 KiB Raw Blame History

Update to 0.13!

2.2 KiB

Raw Blame History