On Tue, Jan 11, 2005 at 11:25:54AM -0500, Bruce Momjian wrote:
> Tom Lane wrote:
> > The latest bug report reminded me that our standard test suite doesn't
> > exercise pg_dump at all. Please do try to beat on RC5's pg_dump over
> > the next week and see if it does the right things.
> The classic test is dump/reload/dump and compare the dumps. I wonder if
> we could automate that.
That wouldn't have caught the index ownership bug. I just did a
test with the 8.0.0rc4 pg_dump and got identical output even though
the restored indexes had the wrong ownership. I don't oppose the
idea -- I'm just mentioning one case where it wouldn't have worked.
In response to
pgsql-hackers by date
|Next:||From: Elein Mustain||Date: 2005-01-11 18:33:44|
|Subject: Re: IBM releases 500 patents|
|Previous:||From: Merlin Moncure||Date: 2005-01-11 17:11:02|
|Subject: problem with casts dump/restore|