This is about two different problems. In one, all log messages are currently being sent to a log file, which means that ERROR and CRITICAL level error messages may not be noticed until the log is read. In the other, I am redirecting STDERR to /dev/null while playing audio with pyaudio. This suppresses the constant "buffer underrun" messages, but is causing a "Too many files" error that eventually crashes Naomi.
Expected Behavior
Error and Critical messages should still appear on the screen even if log messages are being written to the log file. Playing audio should not crash Naomi.
Steps to Reproduce
Select Wit.ai as your STT engine
Run it without entering a valid token
Naomi appears to run fine, but all attempts to address it are recognized as noise.
Check the log file. It should have critical error messages about the invalid token.
Description
This is about two different problems. In one, all log messages are currently being sent to a log file, which means that ERROR and CRITICAL level error messages may not be noticed until the log is read. In the other, I am redirecting STDERR to /dev/null while playing audio with pyaudio. This suppresses the constant "buffer underrun" messages, but is causing a "Too many files" error that eventually crashes Naomi.
Expected Behavior
Error and Critical messages should still appear on the screen even if log messages are being written to the log file. Playing audio should not crash Naomi.
Steps to Reproduce