v3.2.2

Fixed

• Logging no longer breaks (and floods the console with tracebacks) when the log file's handle is invalidated — e.g. the file lives on a network share or removable drive that reconnects, or the host sleeps and resumes. Previously a plain FileHandler would raise OSError (EINVAL) on every subsequent write, lose the record, and print a full traceback per record to the console. The file handler now transparently reopens the file and retries once on a write failure, so a transient blip resumes logging instead of silently dropping everything after it, and logging.raiseExceptions is disabled so an unrecoverable write drops the record quietly rather than dumping a stack trace. Credential redaction is unchanged. (Running the bot from local disk rather than a network share is still the more robust setup.)