Am Montag, 19. Februar 2007 15:57 schrieb Tom Lane:
> The precedent that I'm thinking about is that the command tag for COMMIT
> varies depending on what it actually did.
Some have also argued against that in the past, so I guess we just have
different ideas of how it should work. Not a problem.
--
Peter Eisentraut
http://developer.postgresql.org/~petere/