Bluetooth accessory names with MacRoman apostrophes render as �

[yuchenzhu-research]

Description

Bluetooth accessory names containing a typographic apostrophe can be corrupted in the TUI.

For example:

• Expected: yuchen’s Magic Keyboard
• Actual: yuchen�s Magic Keyboard

The raw output from pmset -g accps can contain byte 0xD5, which represents U+2019 RIGHT SINGLE QUOTATION MARK in MacRoman. Decoding the output with String::from_utf8_lossy treats that byte as invalid UTF-8 and inserts U+FFFD.

Current code

Some(o) if o.status.success() => {
    String::from_utf8_lossy(&o.stdout).to_string()
}

Fix

Keep the command output as bytes, preserve valid UTF-8, and fall back to MacRoman decoding only when needed:

Some(o) if o.status.success() => o.stdout,

fn parse_bluetooth_devices(output: &[u8]) -> Vec<BluetoothDevice> {
    let output = match std::str::from_utf8(output) {
        Ok(text) => std::borrow::Cow::Borrowed(text),
        Err(_) => {
            let (text, _, _) = encoding_rs::MACINTOSH.decode(output);
            text
        }
    };

    // Existing Bluetooth device parsing continues here.
}

Regression tests should cover both MacRoman and already-valid UTF-8 device names.

Pull request

A fix has already been submitted in #18.