Embedded Discovery Service and SP history cookie

[Rod Widdowson]

I just checked, the EDS sets '+' and looks for only that.  I remember having issues in here when I developed this.

AFAICS at this distance, this is because java.net.URLDecoder.URLDecode converts '+' into ' ' and vice versa, 
But the javascript function decodeURIComponent doesn't (and, to my chargrin I tested CDS/EDS cookie interop 
rather more that I did SP/EDS) 

I have entered https://issues.shibboleth.net/jira/browse/EDS-30 to handle this case.



> -----Original Message-----
> From: users-bounces at shibboleth.net [mailto:users-bounces at shibboleth.net] On Behalf Of Takeshi
> NISHIMURA
> Sent: 22 October 2011 20:40
> To: Shib Users
> Subject: Re: Embedded Discovery Service and SP history cookie
> 
> Thanks Scott! I'll dive into the EDS code.
> 
> On 2011/10/23, at 4:12, Cantor, Scott wrote:
> 
> > On 10/22/11 2:44 PM, "Takeshi NISHIMURA" <takeshi at nii.ac.jp> wrote:
> >>
> >> I'm afraid SP and EDS are not compatible with each other if _saml_idp
> >> contains multiple entityIDs. There are two points of view:
> >> 1) Which one is the most recent
> >> 2) What string can be used as a delimiter
> >
> >> SP: the last is most recent, concatenated by '%20'
> >> EDS: the first is most recent, concatenated by '+'
> >>
> >> I don't know which semantics is right, but it is useful for me if both
> >> softwares use the same semantics.
> >
> > The last is the most recent. The encoding shouldn't matter, it's URL
> > encoded, and a + is generally the same as %20 unless there's a bug in the
> > decoder.
> >
> > -- Scott
> 
> --
> To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net