On Thu, Nov 07, 2002 at 12:27:05AM -0500, Tom Lane wrote:
> CLUSTER ALL patch: I have a problem with this, specifically the fact
> that it changes CLUSTER into a multi-transaction operation.
That was your suggestion...
> That renders CLUSTER non-rollbackable and not callable from functions.
> After all the work we went to to make CLUSTER rollbackable, this seems
> like a giant step backward.
Well, CLUSTER ALL is now non-rollbackable. But why is it useful to
rollback a CLUSTER operation?
I think I can make the one-table-only version rollbackable again (and
keep the ALL version multitransaction). Is that a good tradeoff? Note
that the clusterdb script to appear in 7.3 is horribly broken for
concurrent cases, and is much worse than the outstanding CLUSTER ALL
patch.
--
Alvaro Herrera (<alvherre[a]dcc.uchile.cl>)
"Saca el libro que tu religion considere como el indicado para encontrar la
oracion que traiga paz a tu alma. Luego rebootea el computador
y ve si funciona" (Carlos Duclos)
