Noteworthy timestamp change in V4

[Cantor, Scott]

It seems as though the date/time parsing in V3 may have been permissive dealing with timestamps without the trailing Z and handling them as UTC time, but V4 is treating them as some kind of "local" value, not sure exactly what it's doing yet.

Technically that's perfectly correct since it's invalid to omit the Z in SAML, but I do wonder offhand if there are SPs lurking out there generating invalid messages that are going to suddenly break.

I don't think we should try and "fix" this, but it's notable.

Outside of a broken ECP testing script producing Z-less values I have not personally run into any failing SPs in my testing yet.

-- Scott