>>>>> "MW" == Mark Woodward <pgsql@mohawksoft.com> writes:
MW> Yea, I've been toying with the idea of that setting lately, I
MW> can't for the life of me understand why it isn't the default
MW> behavior.
Lots of programs handle malloc() failures very badly. Including
daemons. Often it's better in practice to just keep going a little
longer and see if you can squeeze by -- and then perhaps kill the
memory hog, rather than some daemon getting a NULL and crashing.
/Benny