Closed mtorromeo closed 2 years ago
I just noticed that creating an EventLoop
in create_headless_display
makes it panic when used as a library, but I am not sure how else I am supposed to build a headless display.
Oh hell yes thank you!
And crap, of course something else breaks. The joys of OpenGL -_-. I'll try to figure out what we can do about the EventLoop
as soon as I get the chance.
Also feel free to run cargo fmt
(I know it needs it) or upgrade more crates. Just submit them as separate PRs.
Finally got around to looking at this. This is the error I'm getting when running cargo test
. Is this the same thing you were seeing?
thread 'tests::cube' panicked at 'Initializing the event loop outside of the main thread is a significant cross-platform compatibility hazard. If you really, absolutely need to create an EventLoop on a different thread, please use the `EventLoopExtUnix::new_any_thread` function.', /home/neo/.cargo/registry/src/github.com-1ecc6299db9ec823/winit-0.25.0/src/platform_impl/linux/mod.rs:751:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
I tried the suggested workaround and it works on Linux, but it does indeed stop stl-thumb from building on Windows.
Yes I had a similar error, not sure if it was identical or not.
There seem to be a similar function for windows EventLoopExtWindows::new_any_thread
https://github.com/rust-windowing/winit/blob/1b3b82a3c1369c5248a2e6a251230ba6c615e918/src/platform/windows.rs#L27
Good find. I'll see if I can make that work.
By updating the glium dependency, I managed to solve the segmentation faults with release builds.
This should fix at least #41 and #38, possibly more.
I could have updated more dependencies, but wanted to keep the PR with only relevant changes.
I also had to resist the urge to run
cargo fmt
on the codebase :P