Re: [HACKERS] Interval aggregate regression failure

I am unclear about this report. The patch was not meant to fix every
interval issue, but merely to improve multiplication and division
computations. Does it do that? I think the 23:60 is a time rounding
issue that isn't covered in this patch. I am not against fixing it, but
does the submitted patch improve things or not? Given we are
post-feature freeze, we don't have time to fix all the interval issues.

---------------------------------------------------------------------------

Michael Glaesemann wrote:
>
> On Sep 1, 2006, at 5:05 , Bruce Momjian wrote:
>
> > Tom Lane wrote:
> >> Bruce Momjian <bruce(at)momjian(dot)us> writes:
> >>> Well, the patch only multiplies by 30, so the interval would have to
> >>> span +5 million years to overflow. I don't see any reason to add
> >>> rounding until we get an actual query that needs it
> >>
> >> Have you tried your patch against the various cases that have been
> >> discussed in the past? In particular there were several distinct
> >> examples of this behavior posted at the beginning of the thread, and
> >> I'd not assume that a fix for one handles them all.
> >
> > Yes, it fixes all posted examples, except one that displays 23:60. I
> > cannot reproduce that failure from Powerpc so am waiting for
> > Michael to
> > test it.
>
> Here's your patch tested on my machine, both with and without --
> enable-integer-datetimes. I've tweaked the ad hoc test suite to
> include a case where the days and time differ in sign and added a
> couple of queries to the ad hoc test suite to include the problems
> Tom referred to--not that this patch will fix them, but to keep the
> known problems together. I hope to add more to this to test more edge
> cases.

--
Bruce Momjian bruce(at)momjian(dot)us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +