Re: Updates of SE-PostgreSQL 8.4devel patches (r1268)

Peter Eisentraut <peter_e@gmx.net> writes:

> On Thursday 11 December 2008 18:32:50 Tom Lane wrote:
>> > How can we stick all of these in the same column at the same time?
>>
>> Why would we want to?
>
> Because we want to use SQL-based row access control and SELinux-based row 
> access control at the same time.  Isn't this exactly one of the objections 
> upthread?  Both must be available at the same time.

Well I don't think anyone would actually want them *at the same time*.
Combining multiple security models would mean you aren't actually following
any security model.

But I don't like the idea of making it a compile-time switch. Having to ship
separate packages for different compile-time options is really an awful
solution from the distribution's point of view. And it doesn't scale either --
if we got another such option they would have 2^n combinations.

Distributions like to set distribution-wide policies like "compile with X
support". It doesn't mean you can't run those programs without actually using
that support, as in "emacs -nw". It would be nice to have the option at
run-time of whether to use selinux or row-acl support instead.

I think we need to separate out the --enable-selinux which would merely
compile in the support for selinux from the switch to control whether we
actually have selinux turned on. Make that either an initdb option or a
per-database option like we have with collation/encoding.

Then users can install a single package and decide whether they want to use
selinux or row-acls. If their distribution decides not to compile in selinux
support they just have one choice, row-acls (or nothing).

--  Gregory Stark EnterpriseDB          http://www.enterprisedb.com Get trained by Bruce Momjian - ask me about
EnterpriseDB'sPostgreSQL training!