Re: Bad timestamp external representation

On Wed, Jul 25, 2001 at 06:53:21PM -0400, Bruce Momjian wrote:
> 
> I can confirm that current CVS sources have the same bug.
> 
> > It's a bug in timestamp output.
> > 
> > # select '2001-07-24 15:55:59.999'::timestamp;
> >          ?column?          
> > ---------------------------
> >  2001-07-24 15:55:60.00-04
> > (1 row)
> > 
> > Richard Huxton wrote:
> > > 
> > > From: "tamsin" <tg_mail@bryncadfan.co.uk>
> > > 
> > > > Hi,
> > > >
> > > > Just created a db from a pg_dump file and got this error:
> > > >
> > > > ERROR:  copy: line 602, Bad timestamp external representation '2000-10-03
> > > > 09:01:60.00+00'
> > > >
> > > > I guess its a bad representation because 09:01:60.00+00 is actually 09:02,
> > > > but how could it have got into my database/can I do anything about it?
> > > The
> > > > value must have been inserted by my app via JDBC, I can't insert that
> > > value
> > > > directly via psql.
> > > 
> > > Seem to remember a bug in either pg_dump or timestamp rendering causing
> > > rounding-up problems like this. If no-one else comes up with a definitive
> > > answer, check the list archives. If you're not running the latest release,
> > > check the change-log.

It is not a bug, in general, to generate or accept times like 09:01:60.  
Leap seconds are inserted as the 60th second of a minute.  ANSI C 
defines the range of struct member tm.tm_sec as "seconds after the 
minute [0-61]", inclusive, and strftime format %S as "the second
as a decimal number (00-61)".  A footnote mentions "the range [0-61]
for tm_sec allows for as many as two leap seconds".

This is not to say that pg_dump should misrepresent stored times,
but rather that PG should not reject those misrepresented times as 
being ill-formed.  We were lucky that PG has the bug which causes
it to reject these times, as it led to the other bug in pg_dump being
noticed.

Nathan Myers
ncm@zembu.com