Document - Correct when instantiated with bytes content instead of bytes

[tomdpsrd]

Actual Behavior

Document.summary() is not working with python3 when the document is based on bytes and not on string content.
The new released version (0.8.4.1) contains an old modification that put the regexp in string instead of bytes.

Linked issue :
#194

Steps to Reproduce the Problem

Follow the readme steps

>>>importrequests>>>fromreadabilityimportDocument>>>response=requests.get('http://example.com') >>>doc=Document(response.content) >>>doc.title() Traceback (mostrecentcalllast): ... RE_CHARSET.findall(page) +RE_PRAGMA.findall(page) +RE_XML.findall(page) ^^^^^^^^^^^^^^^^^^^^^^^^TypeError: cannotuseastringpatternonabytes-likeobject