Dmitry Koterov wrote:
> If I use
>
> TRY {
>   EXECUTE SCRIPT ( ... /* CREATE TABLE a(id INTEGER); */ );
>   EXECUTE SCRIPT ( ... /* CREATE TABLE b(id INTEGER); */ );
> }
>
> slonik generates 2 DDL_SCRIPT events which are processed in the above 
> cycle on a replica, and each event is committed in separated transactions.
This is a nice example; thanks!

Yes, if there are multiple events, they won't forcibly get grouped 
together on subsequent nodes even though a TRY{} block might cause them 
to be executed together as a single transaction on the initial node.

I don't see changing that behavior; we'd have to have some rather 
invasive effects on event processing.

I have committed documentation to CVS to indicate this limitation to TRY{}.

-- 
output = ("cbbrowne" "@" "ca.afilias.info")
<http://dba2.int.libertyrms.com/>
Christopher Browne
(416) 673-4124 (land)
"Bother,"  said Pooh,  "Eeyore, ready  two photon  torpedoes  and lock
phasers on the Heffalump, Piglet, meet me in transporter room three"