OpenSAML user discussion

["Scott Cantor" <>]

> I am running into couple issues due to certain characters 
> embedded in the SAML assertion attributes. 
> 
> One is a user with ú  in his first name, and the other is a 
> user with a newline and cr in the postaladdress attribute. 
> Both cases caused signature mismatch between SAML server and receiver.

Signatures are over octet streams, so these sorts of issues mean there are
bugs in the serialization process or something actually got modified.

> We use a different SAML product, but I am curious how 
> opensaml handles this kind of characters.

I use UTF-8 exclusively when writing anything out. It's up to the SAML
application to properly deal with content in the objects that are not ASCII.
Java of course is native Unicode. C++ isn't, but I provide functions to
convert from 16-bit Unicode to UTF-8, but C/C++ can't deal with UTF-8 as
anything but ASCII, so it's a fragile situation.

So to answer the question, I don't deal with them, I ignore them. I don't
have to care that they're there.

-- Scott