On 12/11/2012 10:25:43 PM, Josh Kupershmidt wrote:
> On Tue, Dec 11, 2012 at 11:46 AM, Karl O. Pinc <kop@meme.com> wrote:
> > Yes, the current pg_restore silently
> > ignores multiple --table arguments, and seems to use the last
> > one. You are introducing a backwards incompatible
> > change here.
> Agreed with Robert that this change should be reasonable in a major
> version (i.e. 9.3).
Good by me. Seemed worth a mention.
> >> I believe you need ellipses behind --table in the syntax summaries
> >> of the command reference docs.
>
> Hrm, I was following pg_dump's lead here for the .sgml docs, and
> didn't see anywhere that pg_dump makes the multiple --table syntax
> explicit other than in the explanatory text underneath the option.
Yes. I see. I didn't look at all the command's reference pages
but did happen to look at clusterdb, which does have --table
in the syntax summary. I just checked and you need to fix
clusterdb, reindexdb, and vacuumdb.
> > I also note that the pg_dump --help output says "table(s)" so
> > you probably want to have pg_restore say the same now that it
> > takes multiple tables.
>
> Good catch, will fix, and ditto reindexdb's --index help output. (It
> is possible that the --help output for pg_dump was worded to say
> "table(s)" because one can use a "pattern" --table specification with
> pg_dump, though IMO it's helpful to mention "table(s)" in the --help
> output for the rest of these programs as well, as a little reminder
> to
> the user.)
Agreed.
Karl <kop@meme.com>
Free Software: "You don't pay back, you pay forward." -- Robert A. Heinlein
