Imbris

Hi! I noticed that this depends on `error-chain` which is our only source of a dependency on the `backtrace` crate now that there are methods in std to get backtraces....

winit example works fine

just tried glutin's window example and it works too alt-tab renders properly when I comment out `swap_buffers()` in `Display::draw` in alacritty

yep, alacritty itself isn't rendered (other than what is in in the buffer initially), but the alt-tab interfaces appears and switching to another window renders that window > glutin examples...

I don't have time to investigate further right now, but I guess the next step is cutting out pieces to try to find the difference from the glutin example.

Here is an additional discovery. In addition to alt-tab working when commenting out `self.window.request_redraw()` inside the `display/window.rs`, it also works if I add a request redraw call for every window...

I was able to reproduce in glutin example `cargo r --example window` with these changes: diff ```diff diff --git a/glutin_examples/src/lib.rs b/glutin_examples/src/lib.rs index e5b75e1..2f197fb 100644 --- a/glutin_examples/src/lib.rs +++ b/glutin_examples/src/lib.rs @@ -162,6...

I did some investigation in https://github.com/gfx-rs/wgpu/issues/6046#issuecomment-2309022590 and I believe the trace logs could at least be filtered out more efficiently (when the max filter level is below trace) by passing...

To simplify the example: ``` cargo new cargo-tree-test cd cargo-tree-test cargo add [email protected] --features net cargo tree -e features ``` shows: ``` cargo-tree-test v0.1.0 ├── tokio feature "default" │ └──...

> because "normal" (non-feature) edges are also requested I'm not sure if this is accurate since my example has similar output without `-e normal`. I think including normal just disables...