"Is there some reason why the output of "content" isn't what you really want?"

--actually yes! I'd like to get clean utf8, which requires to first possibly unzip gzipped content and then decode it properly from a local charset to the more universal utf8 representation (and decoded_content should do this in one simple call).