Re: [BUGS] grant/revoke bug with delete/update

[Tom Lane]

Jerome ALET <[EMAIL PROTECTED]> writes:
> Now suppose you define a rule to act differently on delete or update

Hmm, that's the first good argument I've seen in this thread.  An ON
DELETE rule could possibly invoke processing quite different from an
ON UPDATE rule, and in that case it could make sense to want to be
able to allow access to one operation but not the other.

I'm still pretty nervous about trying to make this change at this point
in the 7.0 cycle though, considering that all we have to work from is
an incomplete 6.5.* patch (no docs, no pg_dump fix, even assuming he got
all the backend code changes...).

What do you think?

                        regards, tom lane