Andres Freund <and...@anarazel.de> writes: > Btw, I think Tom's "more that could be done" was referring more to doing > more upfront work, checking, easier input format, whatnot in the genbki, > not so much performance work... Tom, correct me if I'm wrong.
Yeah, as per what we were just saying, performance of that code isn't really an issue right now. Supporting a nicer input format is probably the biggest step forward that could be taken, but that will require some major work :-(. We've also talked about things like supporting regprocedure rather than only regproc (ie disambiguating overloaded functions), likewise regoperator, and so on, with an eye to reducing the number of places where people have to write numeric OIDs in the input. There's some threads on this in the archives, but I'm too lazy to look. regards, tom lane -- Sent via pgsql-hackers mailing list (email@example.com) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers