On Mon, Dec 13, 2021 at 07:53:43PM +0900, Michael Paquier wrote:
> Well, I don't think that it is a big deal one way or the other, as
> we'd finish with InvalidXLogRecPtr for the LSN and 0 for the timestamp
> anyway. If both of you feel that just removing the assertion rather
> than adding an extra check is better, that's fine by me :)
Looked at that today, and done this way. The tests have been extended
a bit more with one ROLLBACK and one ROLLBACK PREPARED, while checking
for the contents decoded.
--
Michael