On Thu, Oct 06, 2022 at 11:51:52PM -0400, Tom Lane wrote:
> Michael Paquier <michael@paquier.xyz> writes:
> > While studying a bit more this thread, I've been reminded of the fact
> > that this would treat different flavors of BEGIN/COMMIT commands (mix
> > of upper/lower characters, etc.) as different entries in
> > pg_stat_statements, and it feels inconsistent to me that we'd begin
> > jumbling the 2PC and savepoint commands with their nodes but not do
> > that for the rest of the commands, even if, as mentioned upthread,
> > applications may not mix grammars.
>
> I've been thinking since the beginning of this thread that there
> was no coherent, defensible rationale being offered for jumbling
> some utility statements and not others.
Only a very small subset causes trouble in real life scenario, but I agree that
cherry-picking some utility statements isn't a great approach.
> I wonder if the answer is to jumble them all. We avoided that
> up to now because it would imply a ton of manual effort and
> future code maintenance ... but now that the backend/nodes/
> infrastructure is largely auto-generated, could we auto-generate
> the jumbling code?
That's a good idea. Naively, it seems doable as the infrastructure in
gen_node_support.pl already supports everything that should be needed (like
per-member annotation).