bin/53899: mktime gives wrong result in Central timezone

[Barry at FreeBSD.org]

>Number:         53899
>Category:       bin
>Synopsis:       mktime gives wrong result in Central timezone
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sun Jun 29 10:10:08 PDT 2003
>Closed-Date:
>Last-Modified:
>Originator:     Barry Pederson
>Release:        FreeBSD 4.8-RELEASE i386
>Organization:
>Environment:
System: FreeBSD eden.barryp.org 4.8-RELEASE FreeBSD 4.8-RELEASE #0: Sat Apr 5 12:35:50 CST 2003 barryp at eden.barryp.org:/usr/obj/usr/src/sys/BARRYP i386


>Description:
	The mktime function seems to give an incorrect result when the timezone
        is set to US Central time.  I first noticed this when running Zope3 (Python)
        unittests, but can also demonstrate the problem using Perl.

>How-To-Repeat:

        Execute mktime() with the same parameters, but vary the TZ enviroment
        variable.  Here is a shell script and a perl script to demonstrate:


------ tzdemo.sh -----------
#!/bin/sh

for tz in 'EST5EDT' 'CST6CDT' 'MST7MDT' 'PST8PDT'
do
    export TZ=$tz;
    /usr/bin/env perl tzdemo.pl $tz
done
----------------------------



------ tzdemo.pl -----------
#!/usr/bin/env perl

use POSIX;

# An arbitrary unix timestamp and its
# gmtime() equivalent.
#

$test_time = 1055176982;
@test_gm = (2, 43, 16, 9, 5, 103, 1, 159, 0);

# You could also say:
#
# @test_gm = gmtime($test_time);

# Convert test time back to a unix timestamp, but since
# mktime() assumes localtime, the result should vary
# according to the current timezone setting
#

$result = mktime(@test_gm);

# Calculate the difference, in hours, between what mktime()
# came up with for a local value  and the initial GMT time
# we started with.
#

$diff = ($result - $test_time) / 3600;

print "$ARGV[0] $result $diff\n";
----------------------------


The expected output is:

  EST5EDT 1055194982 5
  CST6CDT 1055198582 6
  MST7MDT 1055202182 7
  PST8PDT 1055205782 8

The actual output I get is:

  EST5EDT 1055194982 5
  CST6CDT 1055194982 5
  MST7MDT 1055202182 7
  PST8PDT 1055205782 8

The second line (CST6CDT) should not have the same numbers.
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: