> I am about to propose a patch that will cause the default shared_buffers > to be more realistic, say 1000, on machines where the kernel will allow > it. Not sure if people will let me get away with applying it > post-feature-freeze, but if so that would change the terms of this > debate noticeably.
It's not a feature change, it's a bug fix -- bug being an oversight.
signature.asc
Description: This is a digitally signed message part
