Bug #22512 mysql-test-run.pl sometimes does not set correct timezone
Submitted: 20 Sep 2006 11:31 Modified: 7 Nov 2006 18:18
Reporter: Matthias Leich Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Command-line Clients Severity:S3 (Non-critical)
Version:5.1 OS:
Assigned to: Magnus BlÄudd CPU Architecture:Any

[20 Sep 2006 11:31] Matthias Leich
Description:
The internal clock of my box fits to UTC.
Standard timezone within my working environment is
CEST = UTC + 2 hours.
mysql-test-run.pl sets at least for some tests
    lib/mtr_cases.pl:  $tinfo->{'timezone'}= "GMT-3";
./mysql-test-run.pl alias ctype_latin1_de ps_2myisam
TEST                           RESULT         TIME (ms)
-------------------------------------------------------
....
alias                          [ pass ]            259
ctype_latin1_de                [ pass ]            873
ps_2myisam                     [ pass ]           2271
-------------------------------------------------------
Stopping All Servers
Note: ctype_latin1_de has a ctype_latin1_de-master.opt file.
      That means this testcase causes a restart of the server.

./mysql-test-run.pl alias ps_2myisam
TEST                           RESULT         TIME (ms)
-------------------------------------------------------

alias                          [ pass ]            291
ps_2myisam                     [ fail ]

Errors are (from /work/mysql-5.1-falcon/src-3/mysql-test/var/log/mysqltest-time) :
mysqltest: Result content mismatch
(the last lines may be the most important ones)
Below are the diffs between actual and expected results:
-------------------------------------------------------
*** r/ps_2myisam.result 2006-08-16 20:05:20.000000000 +0300
--- r/ps_2myisam.reject 2006-09-20 12:57:08.000000000 +0300
***************
*** 1850,1857 ****
  param08       1991-08-05 01:01:01
  const09       1991-08-05 01:01:01
  param09       1991-08-05 01:01:01
! const10       662680861
! param10       662680861
  const11       1991
  param11       1991
  const12       NULL
--- 1850,1857 ----
  param08       1991-08-05 01:01:01
  const09       1991-08-05 01:01:01
  param09       1991-08-05 01:01:01
! const10       662688061
! param10       662688061
  const11       1991
  param11       1991
  const12       NULL
-------------------------------------------------------

The difference is 7200 = 2 hours.

The bug is most probably not important for customers (-> S3),
but it has a bad impact on server testing and test development
(-> P1).

My environment:
 - Intel PC with Linux(SuSE 10.1)
 - MySQL compiled from source
     mysql-5.1-engines last ChangeSet@1.2336, 2006-09-19
     I guess we have this bug also in other trees and older releases.
     BUILD/compile-pentium-debug-max

How to repeat:
Please execute the commands above.

Suggested fix:
Please make the outcome of
   ./mysql-test-run.pl <any sequence of regression tests>
deterministic = independent of the testing environment.
[7 Nov 2006 18:18] Paul DuBois
Test suite change. No changelog entry needed.