On Monday 12 May 2008 14:40:46 Pavel Stehule wrote:
> 2008/5/12 Tom Lane <tgl@sss.pgh.pa.us>:
> > "Pavel Stehule" <pavel.stehule@gmail.com> writes:
> >> 2008/5/12 Tom Lane <tgl@sss.pgh.pa.us>:
> >>> It would get less annoying if we allowed user-declared exception names.
> >>
> >> Tom, it's exactly like my patch that you rejected two years ago.
> >
> > Uh, no, not "exactly like" --- that patch doesn't have anything to do
> > with the SQL/PSM syntax, and not much with the SQL/PSM semantics.
> > As I read the spec, a condition name isn't a variable and so you can't
> > do runtime assignment to it (and unlike Neil, I don't think you should
> > be able to do so).
>
> In plpgsql I prefer PL/SQL syntax. Mix SQL/PSM and PL/SQL will be
> mismas. But I like idea, so you can set dynamically SQLSTATE and other
> params - because you can write own wrapper for RAISE statement. It's
> can be usable for centralized exception management. I can do it in C,
> but there are lot of users, that could use only plpgsql.
>
I think nod's toward PL/SQL compatability should be given in general. If
people want a PSM style language, let's work on getting pl/psm better
maintained or integrated.
--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL