I just hacked up test018 in RE24 to use a 2.3 consumer, and the final comparison fails. It's mostly benign, the entryUUID occurs in a different position in the 2.3 output. But it looks like 2.3 is doing something funny with the (longer) 2.4 contextCSN. I'll have to play with this some more.
Haven't tried the reverse yet either.