strange behaviour of gettimeofday() and pg_dump

From: Radoslaw Stachowiak <radek(at)alter(dot)pl>
To: pgsql-admin(at)postgresql(dot)org
Cc: pgsql-devel(at)postgresql(dot)org
Subject: strange behaviour of gettimeofday() and pg_dump
Date: 2001-11-15 17:35:00
Message-ID: 20011115183500.T7681@blue.alter.pl
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

OK, once more post about the subject, because after some investigation
it looks like a real and serious postgres bug. (7.1.3)

table definition with one (from many) columns:

timestamp | timestamp with time zone | not null default timeofday()

and the query:
db=# select timestamp,extract( seconds from timestamp) from history where id=274259;
timestamp | date_part
---------------------------+-----------
2001-09-22 20:07:60.00+02 | 59.998003
(1 row)

as You can see, first alement shows 60.00 which is WRONG!
at first sight i thought its average, but its not -> 60.00 !!! (two zeroes after decimal period)

it causes failures at pg_dump and shows that postgres has broken timestamp type.

or maybe i don't understand sth ?

--
radoslaw.stachowiak.........................................http://alter.pl/

Browse pgsql-admin by date

  From Date Subject
Next Message Bruce Momjian 2001-11-15 18:26:50 Re: bug?? in oid2name
Previous Message Luis Amigo 2001-11-15 17:05:17 Re: Importing large data files into PostgreSQL