Hans-Jürgen Schönig <postgres@cybertec.at> writes:
> In my very personal opinion (don't cut my head off) I'd vote for both
> syntaxes.
I'm not opposed to that, although it would be a good idea to check that
Oracle doesn't have some patent covering their syntax.
However, if we go for that then what we probably want to do is implement
the SQL-spec syntax and then add something to translate the Oracle
syntax into a SQL parsetree. We shouldn't need two implementations
in the guts of the system, and I'd expect that translating in the other
direction (SQL WITH to an Oracle internal implementation) wouldn't work,
because WITH does more.
I dunno whether the patch mentioned earlier in this thread could serve
as a starting point for that or not.
regards, tom lane