Improve Error Handling for Dynamic Device Connectivity

[ruffsl]

Summary:
While gilrs handles device reconnects well from a virtual device perspective (retaining gamepad IDs and seamless recovery), the current version of gilrs used in CtrlAssist spams persistent error messages to stdout even after a controller is reconnected. This results in noisy logs and can obscure real issues.

Expected Behavior:
After a controller is unplugged and reconnected, error messages about the lost device should stop once connectivity is restored. Only relevant or actionable errors should be logged.

Current Behavior:
After a disconnect/reconnect event, gilrs continues to emit repeated errors such as:

[2025-12-07T03:26:22Z ERROR gilrs_core::platform::platform::ff] Failed to modify effect of gamepad File { fd: 7, path: "/dev/input/event19 (deleted)", read: false, write: true }, error: ENODEV: No such device
[2025-12-07T03:26:22Z ERROR gilrs_core::platform::platform::ff] Failed to modify effect of gamepad File { fd: 28, path: "/dev/input/event19 (deleted)", read: false, write: true }, error: ENODEV: No such device
...

These errors persist even after the device is reconnected and working.

Possible Cause:

• gilrs may not be fully cleaning up or resetting internal state for force feedback or device handles after a disconnect.
• Log filtering can suppress some messages, but this does not address the root cause.

Suggested Solution:

• Investigate gilrs and underlying library (libudev, evdev) handling of device disconnect/reconnect.
• Check gilrs issue tracker and/or contact maintainers for similar reports or advice.
• Patch log filters as a temporary workaround, but aim for a robust fix that stops error spam after recovery. E.g:

CtrlAssist/src/log_setup.rs

Lines 10 to 17 in b181bcf

	if record.level() == Level::Error
	&& record
	.target()
	.starts_with("gilrs_core::platform::platform::gamepad")
	&& record.args().to_string().contains("epoll failed: EINTR")
	{
	return;
	}

Notes:

• Robust error handling is important for long-running sessions and user experience.
• Consider contributing upstream if a fix is found.