Peter Eisentraut wrote:
> On Thursday 11 December 2008 20:32:25 Tom Lane wrote:
>> Well, the objection I was raising is that they should control the same
>> thing. Otherwise we are simply inventing an invasive, high-cost,
>> nonstandard(*) feature that we have had zero field demand for.
>
> There is certainly a rather big field demand for row-level security. I'm not
> sure about SELinux integration, though, or which one of the two you were
> referring to.
SELinux integration is also strong demand, no need to say.
> The trick, of course, is to make it work well. That would usually require the
> polyinstantiation approach, and I am disappointed that that was apparently
> not chosen here.
As we discussed before, I don't have a plan to implement polyinstantiation
feature. It is too much requirement for enterprise class system, and can be
estimate its implementation so complex.
(Please note that major commercial RDBMSs don't implement it also.)
Thanks,
--
OSS Platform Development Division, NEC
KaiGai Kohei <kaigai@ak.jp.nec.com>