Re: [ILUG] Seconds to date?

From: David Neary (dneary at domain wanadoo.fr)
Date: Mon 02 Sep 2002 - 11:08:00 IST


John P. Looney wrote:
> On Sun, Sep 01, 2002 at 04:14:17PM +0100, Paul Jakma mentioned:
> > > gnu date is limited by time_t. but i thought time_t expired in 2037?
> > (2^31-1)/3600/24/365+1970
> > 2038
> >
> > course, on UltraSparc, x86-64, IA64, alpha, etc:
> > (2^63-1)/3600/24/365+1970
> > 292471210647
>
> May I assume that x86-64 will be able to use a 64bit time_t too?

There's no requirement to use an integer type for time_t at all.
People could already be using a 64-bit time_t on a 32-bit
platform if they wanted (say
typedef struct{
  char block[8];
} time_t;

Or even
typedef double time_t;

In fact I know of people who basically had to re-implement the
standard time functions using a user-defined 64 bit time_t
instead of using the standard library one. In general time_ts are
only supposed to be written and read by the standard time functions
anyway (afaik) - so we could all already be using 64 bit time_ts
if there wasn't so much dependence on time_t being an integer
type.

Cheers,
Dave.

-- 
       David Neary,
    Marseille, France
  E-Mail: bolsh at domain gimp.org


This archive was generated by hypermail 2.1.6 : Thu 06 Feb 2003 - 13:18:35 GMT