Closed GoogleCodeExporter closed 8 years ago
Original comment by andrewar...@gmail.com
on 14 Apr 2009 at 4:21
In fact openid.pl is returning faulty HTTP responses. Here are the headers
from one
such response from openid.pl
Connection: Keep-Alive
Status: 200
Content-Length: 336
Cache-Control: no-cache
Content-Type: charset=utf-8
Date: Tue, 14 Apr 2009 16:21:45 GMT
Set-Cookie: _session_id=9ad2261e817be292b9a7f7d5537ea311; domain=.openid.pl;
path=/
Server: Mongrel 1.0.1
Via: 1.1 RED-PRXY-29
Note that the Content-Type header has an invalid formed value. See the
following
relevant sections of RFC 2616:
http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.17
http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.7
Specifically, Content-Type MUST always have a "type/subtype" value, with an
optional
";parameter" notation, where the parameter might be ";charset=utf-8". So the
HTTP
header is malformed because it lacks the "type/subtype;" prefix.
DotNetOpenId is catching this and rejecting this. A future version of this
library
MAY want to be more accommodating and just quietly ignore faulty HTTP headers,
but
the openid.pl host should be corrected.
Original comment by andrewar...@gmail.com
on 14 Apr 2009 at 4:30
I have emailed marcin.jagodzinski@gmail.com, who is listed at openid.pl as the
contact for that site with the details of what they can do to fix this.
Original comment by andrewar...@gmail.com
on 14 Apr 2009 at 4:41
Thanks a lot for quick answer for this issue.
I also e-mailed Marcin about this case, because some time ago he was looking at
this
problem and these days it looked like problem was on the
stackoverflow/dotnetopenid
side.
Once again thanks a lot.
Original comment by urs...@gmail.com
on 14 Apr 2009 at 4:47
You're welcome, and thanks for reporting it here.
I've modified the library to log an error but still allow OpenID logins to work
when
this occurs as well. It should still certainly be fixed on the Provider-side,
both
because it's the right thing to do, and because plenty of existing RPs out
there will
have the same problem until this is fixed. But future versions of DNOI/DNOA
will be
more tolerant of OPs with this problem.
v2.5 28cd7ff29ec
v3.0 5cf3a29accf
Original comment by andrewar...@gmail.com
on 14 Apr 2009 at 5:20
Original issue reported on code.google.com by
urs...@gmail.com
on 14 Apr 2009 at 3:30