Make sure JSON string values come through consistently as unicode.

[thruflo]

If you look at the current impl of util.extract_json_data it passes request.body directly to simplejson.loads.

Under PY2, when passed a string, loads has the documented behaviour of varying the type of returned string values depending on the content of the strings:

If s is a str then decoded JSON strings that contain only ASCII characters may be parsed as str for performance and memory reasons. If your code expects only unicode the appropriate solution is decode s to unicode prior to calling loads.

You can see the behaviour with e.g.:

>>> import simplejson
>>> simplejson.loads('{"foo":"ear"}')
{'foo': 'ear'}
>>> simplejson.loads('{"foo":"€ar"}')
{'foo': u'\u20acar'}
>>> simplejson.loads(u'{"foo":"ear"}')
{u'foo': u'ear'}
>>> simplejson.loads(u'{"foo":"€ar"}')
{u'foo': u'\u20acar'}

The commit attached follows the advice of "decode s to unicode prior to calling loads" (when running under PY2 and when request.body is a str).

[leplatrem]

Thanks for this PR @thruflo ! The changes look good to me, as far as I understand. Let's ask feedback from our unicode champion @Natim :)

Also do you think you could rebase it to the new master please? (tests will be green)

Thanks again!

[Natim]

Just a nitpick but it looks good to me thanks.

[Natim]

Instead of doing the if not util.PY2 can we use the skipIf decorator here?