Test for 'say' command for TTS fails on macOS #105
Description
In tts.py we check for the availability of the 'say' command for text-to-speech on macOS by trying to run 'say --version'. But say does not support the --version flag, at least not on recent versions of macOS. It returns a non-zero exit status, which makes subprocess.run (with check=True) raise a CalledProcessError exception, which is interpreted as meaning that the say command is not available.
Looks like you can say an empty string. That doesn't make any sound and it returns success.
subprocess.run(["say", ""], capture_output=True, check=True, timeout=5)