Serial write stalls application on some Windows machines about serialport-rs

[bastian2001]

because the read is blocking

That's actually a different issue than mine. This is about the write function. You should be able to use the fork by LukaOber that I linked above. It fixes the read function. In this case, it's the write function that stalls separately from the read function. I've had stalled reads on mine too, but they were resolved when I just used the fork (there's already a merge request for this fork).

[bastian2001]

For different reasons, I reinstalled Windows on my PC. Since then, it works. I don't think there was a bug fix related to this, so AFAIK it's not fixed

[bastian2001]

To give some context: this is my code, with different timeouts specified, all of which stalling the program:

use serialport::new;
use std::io;

fn main() {
    let mut port = new("COM8", 38400)
        .timeout(std::time::Duration::from_millis(100))
        .open()
        .expect("Failed to open port");
    let mut serial_buf: Vec<u8> = vec![0; 100];
    loop {
        println!("Reading");
        let mut s = String::new();
        let res = io::stdin().read_line(&mut s);
        match res {
            Ok(_) => {
                println!("Read {} bytes from stdin", s.len());
                println!("Content: {}", s);
                let s = s.as_bytes();
                port.write(s).expect("Write failed");
                //^^^^^^^^ stalls here
                println!("Wrote {} bytes to Serial", s.len());
            }
            Err(e) => eprintln!("{:?}", e),
        }
    }
}

[FrankLaterza]

I'm having the same issue. I have two threads for reading and writing using a clone of the serial port. I can't write anything on the port because the read is blocking and waiting for a response. This is not an issue on my Linux machine.

[FrankLaterza]

After I changed my crate to the fork by LukaOber it worked great! I had to change some code, but this solution works for me. Thanks for sharing!

[Resonanz]

So.... is this fixed now ?

[nskartoredjo]

I've encountered a similar issue, and with a bit of digging, it turns out that the write timeout is never set for Windows machines. I've created a pull request to resolve the issue