[Httplib2-discuss] Digest authentication buglet?

[Manuzhai <man...@gm...>]

Hello there,

I've been using httplib2 for some things, and it's really nice. I used
to use mechanize and it's derivative mechanoid, and httplib2 feels
much better to me. One thing I would like is a cookie store in the
Http-objects though: so I can really conduct a browser-like session.

The reason I'm writing this email, though, is that I was wondering
about something I walked up against today while trying to authenticate
against the NewsGator REST API. It supports Digest authentication, but
when GETting a response with the DigestAuthentication thing on it, it
returns neither a Authentication-Info nor a WWW-Authenticate header. I
quickly skimmed the RFC, and it seems this behavior is allowed.
However, the DigestAuthentication.response() method, when no
Authentication-Info header is present, tries to parse the
WWW-Authenticate header, which is not actually there (since we've
already authenticated). Therefore, an Exception occurs in
DigestAuthentication.response()...

>>> h, rsp = http.request(URL, headers = head)
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "build/bdist.linux-i686/egg/httplib2/__init__.py", line 764, in request
  File "build/bdist.linux-i686/egg/httplib2/__init__.py", line 616, in _request
  File "build/bdist.linux-i686/egg/httplib2/__init__.py", line 359, in response
KeyError: 'digest'

I hope this can be fixed. If httplib2's behavior is right, I'll gladly
complain to the NewsGator team, but for now it seems this is a problem
with httplib2.

Regards,

Manuzhai