> Is that test good enough to trigger the original bug? In my experience, I had to add a lot more tuples before the logical_decoding_work_mem threshold was crossed and the streaming kicked in. I would suggest running the test without the fix and check if the assertion hits. If so, we are good to go. > Yes, it is reproducing without fix, I already tested it. Basically, I am using the "stream_test" table in "copy stream_test to stdout"" command which already has 20 toasted tuples, each of size 6000 bytes so that is big enough to cross 64kB.