IdP 3.3.1 to 3.4.6 -- transientID difference?

[Cantor, Scott]

> > I guess my question is:  should transientID be different between
> > 3.3.1 and 3.4.6?

Contradicting myself, no, it didn't change in any major way, it's been generated with the cryptographically generated strategy since 3.0.0. If it changed visibly in any versions, it was a change to the internals without any particular relevance to anything except the IdP code internally if it ever decodes them.

The confusion in my memory is that I had considered making the default version Stored even though it requires state because if there are no attribute queries coming in, it just doesn't matter, and a memory-backed approach is faster, with shorter values in the log. I use it myself for that reason.

-- Scott