Use u8 for buffer mapping

[jacob-greenfield]

cc @kvark @grovesNL

This is a temporary solution for #119, and a follow-up for #126.

[jacob-greenfield]

You're on your own for testing multithreaded_compute... that thing butchered my poor laptop 😄

[kvark]

You're on your own for testing multithreaded_compute... that thing butchered my poor laptop

That deserves a separate bug, it's not expected!

[kvark]

This is great, thank you! I only have one naming suggestion

[kvark]

why is this type changed?

[jacob-greenfield]

To get rid of this:

wgpu-rs/src/lib.rs

Line 925 in a5921da

slice::from_raw_parts(data as *const u8, usize::try_from(user_data.size).unwrap())

BufferMapXxxAsyncUserData is only used to pass info to the C callback in the respective map_xxx_async, and the input to those functions is already a usize. The current code converts to BufferAddress and back to usize.

[kvark]

Bors r+

…

On Nov 21, 2019, at 20:36, Jacob Greenfield ***@***.***> wrote: @Coder-256 commented on this pull request. In src/lib.rs: > CreateBufferMapped { id, data } } + /// Creates a new buffer, maps it into host-visible memory, copies data from the given slice, + /// and finally unmaps it, returning a [`Buffer`]. + pub fn create_buffer_mapped_from(&self, data: &[u8], usage: BufferUsage) -> Buffer { Done. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[bors]

Build succeeded

• continuous-integration/travis-ci/push