Re: Updates of SE-PostgreSQL 8.4devel patches (r1704)

On Mon, 2009-03-09 at 16:39 -0400, Tom Lane wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> > On Mon, Mar 9, 2009 at 4:04 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >> Now it's not really KaiGai-san's fault;
> >> the fundamental problem IMHO is that no one else is taking very much
> >> interest in the patch. But that in itself speaks volumes about whether
> >> we actually want this patch or should accept it.
>
> > Are you sure that this isn't just because the original patch was so
> > enormous? If you're referring to reviewing, it's certainly easier to
> > find someone willing to review a 100-line patch than it is to find
> > someone willing to review a 10,000-line patch.
>
> Well, the huge size of the original patch didn't help any, for sure.
> But the nature of this type of problem --- particularly given the
> not-designed-for-it architecture we have --- is that there are going to
> be a lot of subtle issues and very probably a lot of places to touch.
> It gets even worse if you want to put performance constraints on the
> result. I will not have any confidence in SEPostgres until both the
> design and the code details have been reviewed by a fair number of
> experienced PG hackers; and what I see happening is that there simply
> aren't enough of them who care.
>
> If it were a small localized patch I might not particularly care ...
> but what I'm afraid of is that we'll have a monstrous patch that does
> severe damage to readability and modifiability of the backend, and
> has a bunch of bugs to boot (every one of which will qualify as a
> security issue when it's discovered). And on top of that, I'm still
> not sold that there is enough of a user base for it to justify the
> effort we'll have to put into it. If there were, we'd be seeing more
> interest in reviewing it.

Can't it be kept separately maintained release for a version or two, so
that we will have both PostgreSQL and SE-PostgreSQL and thus have an
easy way to compare both correctness and performance ?

Anyone remember how did Linux implement/introduce SE Linux ?

--
Hannu Krosing http://www.2ndQuadrant.com
PostgreSQL Scalability and Availability
Services, Consulting and Training