Re: [PATCH] remove redundant ownership checks

On Thu, Jan 14, 2010 at 1:04 AM, Greg Smith <greg(at)2ndquadrant(dot)com> wrote:
> But the preference of the last CF is to not apply any patch which doesn't
> have a very clear justification to be committed.  Given that whether this
> patch is applied or not to 8.5 really doesn't make any functional
> difference, I don't see anywhere for this to go right now except for
> "Returned with Feedback".  It's extremely valuable to have had this patch
> submitted.  I don't believe an exact spot of contention with the current
> code was ever highlighted so clearly before this discussion, because
> previous patches were just too big.  We do need to get this whole thing off
> the list for a while now though, I think it's gotten quite a fair slice of
> discussion already.

While I understand your desire to get this patch closed out and move
on to other things, I don't agree that we've given any meaningful
feedback as yet. We really haven't made any progress getting an
agreement on where or how the permissions checks should be done. I
would be perfectly happy to throw this patch under the bus in exchange
for some meaningful feedback on what a more comprehensive approach
would look like, but so far none has been forthcoming - and not
because it hasn't been requested.

http://archives.postgresql.org/pgsql-hackers/2009-12/msg01824.php

It is my view that no patch which makes substantial changes to the
security checks in the code is likely to get committed without Tom's
approval. If Tom is not willing to provide input on a comprehensive
plan, and is also not willing to accept even the least-consequential
change without a fully-fleshed-out comprehensive plan, then I think we
are at an impasse. Note that I am NOT saying it is Tom's
RESPONSIBILITY to provide input on a comprehensive plan. I am only
expressing my opinion that no progress can be made otherwise.

...Robert