On Thu, 16 Apr 2020 at 21:16, Shay Rojansky <r...@roji.org> wrote: > Npgsql would be fine. In fact, Npgsql doesn't have any specific expectations > nor any specific logic around commit; it assumes errors may be returned for > any command (COMMIT or otherwise), and surfaces those errors as .NET > exceptions.
Hi all, I work on the pg8000 Python driver for Postgres and having read through the thread I'd like to echo Shay Rojansky's comment and say that pg8000 would be able to handle the behaviour resulting from the proposed patch and I support the change of a call to commit() *always* producing an error if it has failed. I can understand people's reluctance in general to change server behaviour, but in this case I think the good outweighs the bad. I think most people expected the server to be behaving like this anyway. Regards, Tony.
