All exceptions are caught

Tue to the new structure of scan_base, all exceptions happening in _configure, _scan or _analyze are caught and logged only via logging.error(e) which only gives you the error message. The exception type and traceback are lost forever. This makes developing scans near impossible.

I think it only makes sense for very special cases to catch an exception during one scan of one chip and continue with the other chips. The normal case should be to raise an exception immediately. At the very least, we should save all occurred exceptions and re-raise them after the scan has finished.

What's your opinion?