On Tue, Sep 08, 2020 at 12:18:23PM -0700, Andres Freund wrote: > A test verifying that a non-transactional message in later aborted > transaction is handled correctly would be good.
On top of that, the patch needs a rebase as it visibly fails to apply, per the CF bot. -- Michael
Where are you with this? Are you able to work on it ?
Dave Cramer
Apologies for the delay, here.
I've attached v2 of this patch which applies cleanly to master. The patch also now includes a test demonstrating that pg_logical_emit_message correctly sends non-transactional messages when called inside a transaction that is rolled back. (Thank you, Andres, for this suggestion.) The only other change is that I added this new message type into the LogicalRepMsgType enum added earlier this week.