Wgpu-rs: Panic on running capture and hello-compute examples

[imjasonmiller]

Hi,

I'm wondering whether this might be something on my end. Running the mentioned examples on b1f3fba results in a panic on my system.

I'm currently running Wayland on Arch Linux 5.6.15 with sway as a window manager.

$ cargo run --example capture
$ cargo run --example hello-compute 1 2 3 4

Both result in the following panic message:

thread 'main' panicked at 'You cannot unmap a buffer that still has accessible mapped views', src/lib.rs:379:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
thread 'main' panicked at 'unable to remove range from map context', src/lib.rs:411:21

Commenting out output_buffer.unmap() in the capture example allows it to not crash.

[kvark]

@lachlansneff could be related to your latest mapping changes?

[cwfitzgerald]

Adding a

drop(padded_buffer)

or not manually unmapping at capture/main:156 solves htis issue.

[lachlansneff]

Oops, this looks related!