Instead of returning "malformed section" on trailing bytes or truncated sections, attempt to return parsed results. This handles older HRRR files (2019) that have padding bytes after the last section.