David Fetter <david@fetter.org> writes:
> Presumably psql should know about this change. Should \df now include
> windowing functions along with a boolean column that indicates whether
> a function is a windowing function? Should there be \dw[+] instead?
> In either case, should the S option indicating "include system
> functions only when S is present" (e.g. \dwS) apply?
If people are going to start proposing that, I'm going to switch back to
the position that WINDOW should be treated as an attribute. The whole
point of the discussion is that no one wanted to get into the game of
treating window functions as a separate classification in general.
regards, tom lane