This is not right.

On Python 2, json.dumps() might return Unicode in some cases. It doesn’t by default; yet we could use to change the default:

data = json.dumps(recent_posts, ensure_ascii=False, indent=2, sort_keys=True)
with io.open(output_path, "w+", encoding="utf-8") as outf:
    try:
        outf.write(data.decode('utf-8'))
    except AttributeError:
        outf.write(data)

Absolutley we can do it like that. Don't know why the first commit is wrong though, isn't json utf-8 by design? In ptyhon2 ".decode('utf-8')" works on both string and unicode objects

It does not work properly.

>>> u"ą".decode('utf-8')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.7/encodings/utf_8.py", line 16, in decode
    return codecs.utf_8_decode(input, errors, True)
UnicodeEncodeError: 'ascii' codec can't encode character u'\u0105' in position 0: ordinal not in range(128)

Please switch it to the solution I recommended.

Actually, in this case we need

except (AttributeError, UnicodeEncodeError, UnicodeDecodeError):

yea, excepting more thing makes it more clear.

We don't decode random unicode, we decode output from json.dumps:

>>> import json
>>> json.dumps([u"ą"]).decode('utf-8')
u'["\\u0105"]'
>>>

Just do it with ensure_ascii=False. More modern.