We need to see the code that you are using to get from your date/time values to the epoch seconds. By my calculations, '31.10.2004 00:40' is 1099179600 - which gives the expected difference of 1200.

Of course, the difference between your answer and the expected answer is one hour, so daylight saving time could be an issue. Which timezone are you in, and when do you change out of daylight saving time? Maybe you could try using 'timegm' instead of 'timelocal' and see what difference that makes.

Here's the code I used.

#!/usr/bin/perl

use Time::Local;

my $t1 = '31.10.2004 01:00';
my $t2 = '31.10.2004 00:40';

$t1 = time2epoch($t1);
$t2 = time2epoch($t2);

print "$t1 - $t2 = ", $t1 - $t2, "\n";

sub time2epoch {
  my $t = shift;

  my ($d, $m, $y, $H, $M) = split /[\.\s:]/, $t;

  --$m;
  $y -= 1900;                                

  return timegm 0, $M, $H, $d, $m, $y;
}
[download]

The duplicated hour in local time during the switch back to "normal" time (from DST) makes the conversion to UTC ambiguous during that period (once a year.)

If you need to avoid that issue you should probably store all dates and times in UTC, and only convert to localtime on display...

Michael

In the GMT timezone the day light saving will have ended in that period. If you want to have consistent calculation do not use timelocal - which will take that into account but use timegm() instead.

/J\

#! /usr/bin/perl

use Date::Calc qw(Delta_DHMS);

($days, $hours, $minutes, $seconds) =
  Delta_DHMS( 2004,10,31,00,40,00,  # earlier
                2004,10,31,01,00,00); # later

print "$days days and - $hours:$minutes:$seconds \n";
[download]