Ideal is to know that once event loop is running, it will not panic.
That is really hard, but can at least identify and reduce panics. At worst, errors should lead to packets being silently ignored. This places a big onus on backend code - it can return error information, but never panic.
Existing code has way too many unchecked optimistic unwrap() calls.
Panics when loading configuration files at start-up may be correct response.
Ideal is to know that once event loop is running, it will not panic.
That is really hard, but can at least identify and reduce panics. At worst, errors should lead to packets being silently ignored. This places a big onus on backend code - it can return error information, but never panic.
Existing code has way too many unchecked optimistic unwrap() calls.
Panics when loading configuration files at start-up may be correct response.