3.2.x and Earlier Login Switching

[Klingenstein, Nate]

> That depends on how screwed up your password authentication processes
> are. We have a half-dozen identifiers they've forced me to accomodate, so
> my input side is a mess. If you don't have that problem, that helps.

Ew.  No.

> Depends where you're accessing that context. You'd have to get it from the
> pool of previous results, but honestly I don't know why you would. Just do it
> the way I did.

The canonicalization class doesn't appear to be included in 3.2.1 as distributed and I need to hew as close to distributed and supported as possible, whether by Shibboleth or the MFA vendor.  That's all.  I could selectively include these classes, but the more I do, the more I have to explain.

In this case, less features are more.