Andrew Gierth <andrew@tao11.riddles.org.uk> writes:
> "Tom" == Tom Lane <tgl@sss.pgh.pa.us> writes:
> Tom> BTW, the reason we need to store the column count explicitly is
> Tom> that we have to ignore the added columns if a composite type has
> Tom> had an ADD COLUMN done to it since the RTE was made. The
> Tom> submitted patch fails rather nastily in such cases, if the
> Tom> composite type isn't last in the function list.
> Am I understanding correctly that the only reason this didn't fail
> before we added ORDINALITY is that the executor in general does not
> care if there are more columns in a tuple than it expects? And that
> adding ORDINALITY broke this already?
Probably it's already broken with ORDINALITY, but I've not checked.
regards, tom lane