Andres Freund <andres@2ndquadrant.com> writes:
> On Saturday, October 20, 2012 12:05:15 AM Tom Lane wrote:
>> (such as the current query showing up in pg_cursors --- maybe we should
>> prevent that?)
> I don't really see an argument for preventing that.
Well, the reason it seems peculiar to me is that the current query is in
no way a cursor --- it's just a SELECT in the cases that showed
regression test differences. I didn't go looking in the code yet, but
I suspect the pg_cursors view is displaying all Portals. Arguably, it
should only display those that were created by actual cursor commands.
regards, tom lane