Yes, but the way you've written it you'd be comparing timestamps stored (directly or indirectly) in the same Session. If you wrote

Assuming that Session exposed this context (it won't, but for the purpose of illustration for this discussion...),

As I explicitly pointed out, Session will not expose this. It will no longer be kept on Session. But even if it were, your point is still not valid. If you actually look at the code Session's current-txn-start-timestamp is that... a timestamp when the current transaction started. And as we have 2 txns here that would mean 2 different values.

I am confused - I thought you use CTC == StorageAccessContext

It is. Again, as I said explicitly:

The reason I asked about those specific assertions is that there is an related option where we can simply reuse the CacheTransactionContext instance, the assumption there being that the CacheTransactionContext itself is aware of the txn boundaries.

You seem to be taking parts of things I say out of context. I clearly said here that CTC would itself be aware of txn boundaries.

what existing option?

No idea what this is in reference to. Where did I mention an "existing option"?