On Thu, 2002-03-07 at 21:21, Tom Lane wrote:
> I dunno. I think the direction we were planning to go in was that -t's
> argument should be treated as a pattern to match against table names
> (eg, a regexp). It's not clear that zero matches are an error when
> you are thinking in those terms. On the other hand I can see your point
> about not realizing your "backup" isn't.
Can't the regex mode be made a different switch (say, -T or -E)? In
addition to allowing better error reporting, I'm uncomfortable with
saying "a table name must consist of non-regex special characters",
which is basically what this implies.
Cheers,
Neil
--
Neil Conway <neilconway@rogers.com>
PGP Key ID: DB3C29FC
