Shibboleth XML parser - issues with parsing dash character

[Ian Young]

> On 28 Nov 2016, at 20:29, Cantor, Scott <cantor.2 at osu.edu> wrote:
> 
>> A SAML trace provided by the external IDP identified that the problematic
>> users had a SAML token that contained an endash character ‘–‘ (Unicode Hex
>> 'x96') in one of the attributes of the SAML token.
> 
> I'm not aware of that being an invalid character in the abstract, but the specific encoding used may have been invalid UTF-8.

If there is a single octet containing 0x96, then that would be an invalid encoding in UTF-8. The leading octet in a UTF-8 encoding MUST NOT have the top two bits set to 10xxxxxx, that combination is reserved for continuation bytes.

The rule is that a UTF-8 encoding is either a single-octet 0xxxxxxx, or a leading 11xxxxxx followed by some number of continuation 10xxxxxxx.

I'd guess this is a mislabelled (or unlabelled) single-character encoding like Windows Cp1252 (that has a U+2013 EN DASH in the 0x96 position).

    -- Ian




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://shibboleth.net/pipermail/users/attachments/20161201/90461abc/attachment.html>