Many thanks to Antti Lampinen for reporting a wrong results issue with merge-right-anti-join. I can't imagine how awful I would feel if this issue had persisted in production for a longer time. postgresql.org/message-id/fla…
Maybe we need to mark Vars nullable by grouping sets, just like we mark Vars nullable by outer joins. Doing that will help fix several long existing semantic bugs in grouping sets. #postgrespostgresql.org/message-id/CAM…
This reminds me of a quote from Tom Lane: (A wise man once said that the most powerful bug-finding heuristic he knew was "where else did we make this same mistake?" ). It's so true.
While working on a crash issue in expandRecordVariable() I found another one that happens in get_name_for_var_field() by grepping the source codes. #postgres