Re: Per-function GUC settings: trickier than it looked

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, Sep 05, 2007 at 02:13:57PM +1000, Brendan Jurd wrote:
> On 9/5/07, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> > "Brendan Jurd" <direvus(at)gmail(dot)com> writes:
> > > Am I on the right page?
> >
> > Got it in one, I believe.
>
> In that case, +1 for your proposed changes.
>
> At first, like Florian, I found the idea of a SET LOCAL ever
> persisting beyond a function astonishing, but that's because I was
> approaching the term LOCAL from a programming frame of mind, not an
> SQL one [...]

As an unqualified POV, seeing that this got at least two people confused
- -- wouldn't it make sense to be more verbose and call the thing SET
TRANSACTION LOCAL (not just TRANSACTION, which is ambiguous as we have
already seen). May be even SET LOCAL TO TRANSACTION (that gives at least
some room for possible extensibility).

I know too little about the parser to have even an idea whether this
would be feasible at all.

Regards
- -- tomás
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFG3mFABcgs9XrR2kYRAug1AJ9FJdFEjDGpYWSj09+LgRv218efdwCcDBR8
kjE8O+QCdD/DMntr6mjHBoA=
=FI+2
-----END PGP SIGNATURE-----