Hi Dan,

On 23 June 2017 at 19:41, Dan <dan15i@yahoo.com> wrote:
Do PWB has issues with decoding URL strings?

Nothing in your example suggests it does:

        test1 = urllib.unquote(m)
        test2 = urllib.unquote_plus(m)
        test3 = m.decode('utf8')
        test4 = m.encode('utf8')

These are all questions of what the Python built-in urllib module does. In the case of Python 2, the behavior is a bit odd, and I think this is what is causing your issue.

In your example, m = u'%c3%85', i.e., a unicode string with the text "%C3%85". Urldecoding this should yield two bytes: the bytes C3 and 85, i.e, the UTF-8 representation of Å.

However, what Python 2 does is it interprets u'%c3%85' to mean 'a unicode string with characters U+00C3 U+0085', i.e., the characters Ã and [unprintable]. There is no clean way to fix the situation after we have ended up there.

Now -- how to solve this?

- The most obvious solution is 'Use Python 3', where the unquote function correctly processes the string.
- Another option is to turn your URL into a bytestring first, i.e., m = m.encode('utf-8'), then call unquote, then decode the string again.
- As you already have a dependency on pywikibot, the last option is to use the pywikibot.page.url2unicode, which works correctly, even on Python 2.