Shib IDP/ECP and input error

[Edwards, Wendy A]

Thanks!  I’m actually using a simpler script at
http://www.cilogon.org/ecp/testecp.sh .  But comparing that script to
Scott’s at 
https://wiki.shibboleth.net/confluence/download/attachments/4358416/ecp.sh
, it does look like the curl call in the testecp.sh script may not be
setting the content-type correctly (as IDP2 would apparently tolerate, but
IDP3 won’t).  I think I’ll try updating the testecp.sh script to set the
content type to "application/soap+xml" and see if that affects the
behavior.

On 4/27/15, 12:23 PM, "Scott Koranda" <skoranda at gmail.com> wrote:

>Hi Wendy,
>
>I just updated the ecp.sh bash script yesterday. If you were using a
>previous version please try with that.
>
>The old version of that script did a POST to the IdP with the
>Content-Type set incorrectly. The v2 IdP was forgiving but the v3 IdP
>"needs" to see the proper Content-Type of "application/soap+xml".
>
>I expect this is actually a client error and not a v3 IdP error.
>
>With my updated script I am able to make ECP work with v3 IdP with no
>problems.
>
>Thanks,
>
>Scott K
>
>P.S. I have not updated the Python version yet. Working on that...
>
>
>On Mon, Apr 27, 2015 at 11:50 AM, Edwards, Wendy A
><wedwards at illinois.edu> wrote:
>> Thanks, Scott.  Done: https://issues.shibboleth.net/jira/browse/IDP-708
>>
>> On 4/27/15, 11:34 AM, "Cantor, Scott" <cantor.2 at osu.edu> wrote:
>>
>>>On 4/27/15, 12:16 PM, "Edwards, Wendy A" <wedwards at illinois.edu> wrote:
>>>>
>>>>However, when I examined the input in the debugger, it was possible to
>>>>find the SOAP XML code (which appeared to be correct) associated with
>>>>the
>>>>request.  Could there be any possibility that an InputStream is being
>>>>reused in the code somewhere?
>>>
>>>I guess, but the decoders all work the same way regardless of what the
>>>client is, so it would be unusual for one to work and not the others.
>>>
>>>It must be a regression because I tested with Scott's shell client
>>>originally, though I tested with the basic-auth flow.
>>>
>>>>If this appears to simply be a misconfiguration somewhere on my end,
>>>>I¹d
>>>>be happy to fix it and retry.  If it could be a code issue, I could
>>>>look
>>>>for a workaround and report back.
>>>
>>>I really couldn't say, I'll have to debug it, so please file one.
>>>
>>>-- Scott
>>>
>>>--
>>>To unsubscribe from this list send an email to
>>>dev-unsubscribe at shibboleth.net
>>
>> --
>> To unsubscribe from this list send an email to
>>dev-unsubscribe at shibboleth.net
>-- 
>To unsubscribe from this list send an email to
>dev-unsubscribe at shibboleth.net