Tomas Vondra <tomas.vondra@enterprisedb.com> writes:
> The other aspect I dislike is that we just stop caching publication
> info, forcing us to reload it for every replicated change/row.
Ouch --- that seems likely to be completely horrid. At the very least
I'd want to see some benchmark numbers before concluding we could
live with that.
regards, tom lane
