Transfer encodings (gzip, compress, etc) are already removed automatically. No need to use decoded_content for that.

decoded_content does remove some content encodings. This means, for example, that it allows you to request a .gz file and get the .xml within. But the OP didn't say anything about requesting a non-XML file and wanting to get the XML from it.

Finally, and most relevant, it also undoes character encodings, and that's bad. It can be told not to do this, but my comment was directed a code that didn't prevent this.