logic bug in RequestsUrlReader means read_compressed is always used.

[Scott Moser]

RequestsUrlReader has this blob of code in its __init__
        ce = self.req.headers.get('content-encoding', '').lower()
        if 'gzip' in ce or 'deflate' in ce:
            self._read = self.read_compressed
        else:
            self._read = self.read_raw

That if statement will always take the rad_compressed route
because bool('gzip') will always return true.

So we really needed:

if 'gzip' in ce or 'deflate' in ce:
    ...