To avoid this warning your RP need to publish RP discovery
information; see
http://code.google.com/p/openid4java/wiki/RelyingPartyDiscovery
> The user can continue, but I am not receiving any attributes at all. I
> don't know if these issues are related or if I'm somehow asking for
> the attributes incorrectly.
As far as I know Yahoo does not support SReg or AX. It may be worth
asking them at
openid-...@yahoo-inc.com
Johnny
>> As far as I know Yahoo does not support SReg or AX. It may be worth
>> asking them at
>> openid-feedb...@yahoo-inc.com
>
> What is the point of using openid if all it does is give the user a
> unique identifier that tracks them across sites?
>
> They would still have
> to choose a unique local username as well as enter any required
> information like birth date.
Building reputation, not having to remember lots of passwords, more
secure authentication at the OP (as applicable).
I'm not disagreeing that AX is useful (on the contrary, and
especially with site registration!), just pointing out that OpenID
Authentication does have value in itself.
> It seems to me that using openid without attributes would just cause
> the registration process to take even longer than it would if openid
> wasn't used at all.
>
> I suppose that from the user's point of view it would be a slightly
> longer registration than normal, but they wouldn't need to remember
> more than one password.
The site-specific username (if any) would not have to be remembered
either.
> Dang. I was hoping to encourage more drive-by people to use my sites
> while logged in by using openid. But, if they need to enter
> registration information anyway then it doesn't seem likely to help
> much.
These are all very good issues, and again I recommend that you let
the Yahoo people know your opinion. My impression is that they truly
appreciate feedback.
Johnny
> Yahoo! is doing a GET on my returnto_url. I'm not sure what they are
> looking for there.
They are doing RP dicovery; see:
http://code.google.com/p/openid4java/wiki/RelyingPartyDiscovery
Johnny
> I reviewed that and the spec. I have that information in my XRDS
> document:
>
> $ curl -ik https://test.fluketeam.com/openid?do=xrds
It seems then that Yahoo's Yadis discovery fails on https://
test.fluketeam.com/openid?do=authenticate.
Try using https://test.fluketeam.com/openid as the realm, and then
have the xrds file served directly at this URL. You can also try
letting them know that your apparently proper RP discovery setup
fails their tests.
> I'm a little confused on how Yahoo! discovers where to find the XRDS
> document.
It performs Yadis discovery on the realm or return_to that your RP
sends in auth requests.
"The relying party MAY publish this information at any URL, and
SHOULD publish it under the realm so that providers can verify
return_to URLs."
http://openid.net/specs/openid-authentication-2_0.html#rp_discovery
> I noticed Yahoo! OpenID authentication works when
> authenticating with plaxo.com. Any other examples?
https://verify.sxip.com/papedemo/
Johnny
> Thanks for your help. I contacted Yahoo and they have a tiny bug they
> are already working on a fix for. I simply had to set the content
> type header as the workaround:
>
> response.setContentType("text/html");
> response.setHeader("X-XRDS-Location", "https://test.fluketeam.com/
> openid?do=xrds");
Good to hear that it's working for you.
> And through testing, only the X-XRDS-Location header is needed,
I'd agree that the header is sufficient, but it's not needed (as in
required).
> not:
> <meta http=equiv="x-xrds-location" content="https://
> test.fluketeam.com/
> openid?do=xrds" />
A correct Yadis implementation will take either into account, with
the preference given to the header. Also, text/html content-type for
the XRDS document is acceptable per the Yadis spec.
It's also good to know that Yahoo is working on fixing these glitches.
Johnny