When the x11 server crashes, loop connection.wait_for_event will occupy 100% of a single cpu core

[I-Want-ToBelieve]

wait_for_event should return an error when the x11 server crashes, so that the program can catch the error instead of entering an infinite loop

use breadx::{display::DisplayConnection, prelude::*};

fn main() {
    /*
     * @see https://docs.rs/breadx/3.1.0/breadx/
     */
    let mut connection = DisplayConnection::connect(None).expect("should connect to x11 server");

    /*
     * @see https://gist.github.com/ssokolow/e7c9aae63fb7973e4d64cff969a78ae8
     */

    // primary event loop
    loop {
        let event = connection.wait_for_event();

        match event {
            Err(_) => {
                eprintln!("X11 server has crashed, exiting program.");
                std::process::exit(1);
            }
            Ok(_) => todo!(),
        }
    }
}

I've prepared a minimal repo that can easily reproduce the issue

git clone https://github.com/I-Want-ToBelieve/autohide-tdrop.git
git checkout bug/infinite-loop 
cargo run
pkill -f Xwayland