Am I barking up the wrong tree? Maybe I'm missing something obvious?
I'm running 1.38 of ikiwiki and the newest CGI::Session, Net::OpenID::Consumer,
-Crypt::DH, URI-Fetch. --Ethan
\ No newline at end of file
+Crypt::DH, URI-Fetch. --Ethan
+
+> The url must not have a setting for openid.mode or openid_identifier in
+> it. So the OpenId plugin didn't know that it was trying to log in. I
+> think this points to an issue with the OpenID server. --[[Joey]]
+
+>> I put debugging output in openid.pm and it suggests that the
+>> verification is taking place successfully. I see "openid.mode=id_res"
+>> in the URI. On top of that, it's the same Openid server I use
+>> to sign in here on ikiwiki.info. --Ethan
+
+>>> Yikes, I don't really have the newest CGI::Session after all..
+>>> let me try updating that. --Ethan
+>>>> Sorry, I'm an idiot -- cookies disabled on my browser. Sorry to
+>>>> waste your time.. --Ethan
+
+>>>>> No problem, the error message could certianly use improvement.
+>>>>> Although if I disable cookies, myopenid lets me know. Maybe you
+>>>>> should paste the url. --[[Joey]]
+
+I have cookies disabled on my computer, with a bunch of manual
+exceptions. This includes myopenid, ikiwiki.info, livejournal,
+and some others. Unfortunately it didn't include my own domain.
+So the URI that myopenid redirected me to was fine, but because
+I didn't have cookies set, I didn't have a session, and so
+session->param('postsignin') was undefined, so instead of being
+redirected my query fell through CGI.pm to the bottom of cgi(),
+where I got the message above. In a perfect world I'd say that
+it would be nice to let the user know that they can't sign in
+w/o cookies, but I don't see any easy way of detecting that
+from CGI::Session. Maybe you know a way -- I have never used
+CGI.pm before, this isn't my forte (in case that wasn't obvious).
+--Ethan
+
+> It's not easily possible to test for cookies, but it is possible to
+> display a better error message in this failure mode. [[bugs/done]]
+> --[[Joey]]