in all cuurently known log files 

After that it will by default use log tables. That setting can be overruled
by adding a -master.opt file like mysql_client_test has.

if that is same as opt_mem it can be removed

with the function where vardir's are created

 - Better control of when to create/remove vardir


mysql-test/mysql-test-run.pl:
  Split the part the function cleanup_stale_files in two parts
    - one that removes any old var directories
    - one that setup var in te proper location
  
  If running against extern server and no vardir has been created in the specified
  location, run the latter function to create the vardir structure. Else print
  message saying the vardir already exists and continue.

And if var/ has been created, it will be recreated later so the logs
from "mtr_kill_leftovers" will be lost anyway  


mysql-test/mysql-test-run.pl:
  Don't create vardir before mtr_kill_leftovers
mysql-test/lib/mtr_process.pl:
  Send output - if any - from mysqladmin to stdout/stderr. There shouldn't
  be any such errors and if any occur, it's good to know about them.

mysql-test/lib/mtr_report.pl:
  Don't output "skipped" comment if there is none (bug#24471)
mysql-test/mysql-test-run.pl:
  Removed "use diagnostics", reduces Perl speed significantly. Can be
  enabled with "perl -Mdiagnostics mysql-test-run.pl".

traces in Valgrind (broken libc6-dbg).
Installing libc6-dbg on Debian will still provide proper bactraces, even
without setting LD_LIBRARY_PATH explicitly.


mysql-test/lib/mtr_io.pl:
  Do not link /usr/lib/debug/* on Debian <=3.1, as it causes broken stack
  traces in Valgrind.
mysql-test/mysql-test-run.pl:
  Do not link /usr/lib/debug/* on Debian <=3.1, as it causes broken stack
  traces in Valgrind.

into  mysql.com:/usr/home/ram/work/bug21789/my41-bug21789


mysql-test/r/date_formats.result:
  Auto merged
sql-common/my_time.c:
  Auto merged

into  mysql.com:/usr/home/ram/work/bug22029/my41-bug22029


mysql-test/r/date_formats.result:
  Auto merged
mysql-test/t/date_formats.test:
  Auto merged
sql/item_timefunc.cc:
  Auto merged

into  shellback.(none):/home/msvensson/mysql/mysql-4.1-maint

$MYSQLTEST_VARDIR, for usage of vardir=e:/var<nnn> on windows


client/mysqltest.c:
  Add auto replace for $MYSQLTEST_VARDIR
  Only add strings with strlen > to patterns array - avoids
  lot of strlen calls

into  neptunus.(none):/home/msvensson/mysql/mysql-4.1-maint

Update usage description

mysql-test/mysql-test-run.pl:
  Don't require restart if master is not started when running with --extern

mysql-test/lib/mtr_cases.pl:
  Don't look at lines that start with # ie. a comment
mysql-test/mysql-test-run.pl:
  Set env BIG_TEST if --big-test passed

into  example.com:/work/bug24219/my41

into  neptunus.(none):/home/msvensson/mysql/mysql-4.1-maint


mysql-test/mysql-test-run.pl:
  Manual merge of fix for location of mysql_client_test

into  mysql.com:/usr/home/ram/work/bug23653/my41-bug23653

There was an improper order of doing chained operations.

To the documentor: ENABLE|DISABLE KEYS combined with RENAME TO, and no other
ALTER TABLE clause, leads to server crash independent of the presence of
indices and data in the table.


mysql-test/r/alter_table.result:
  update result
mysql-test/t/alter_table.test:
  add test for bug#24129
sql/sql_table.cc:
  If there is operation on the KEYS, first do it
  and then do a rename if there is such. Or, we will crash because
  the underlying table has changed.

into  zippy.cornsilk.net:/home/cmiller/work/mysql/mysql-4.1-maint

	with embedded spaces
  
Problem:  Original code written assuming password contains no spaces.
Further, because of how sh interprets characters in variables, you
cannot do the "simple solution".
  
Solution:  Move the "was a password provided" check to the end. This
allows us to correctly quote the password argument when passed to the
mysql client program.  Addendum:  Put this check inside of a shell fn
to simplify the code.


scripts/mysql_fix_privilege_tables.sh:
  Moved password existence check to end to allow for passwords with
  with embedded spaces.

into  polly.local:/home/kaa/src/maint/mysql-4.1-maint

mysql-test/mysql-test-run-shell.sh:
  Rename: mysql-test/mysql-test-run.sh -> mysql-test/mysql-test-run-shell.sh
mysql-test/Makefile.am:
  Remove build of mysql_test_run_new
  Symlink mysql-test-run.pl to mysql-test-run
  Rename mysql-test-run.sh to mysql-test-run-shell.sh

use MTR_MEM environment variable

into  zippy.cornsilk.net:/home/cmiller/work/mysql/mysql-4.1-maint

Support says that memlock doesn't work on OSes other than Solaris.  
Add a warning about --memlock to the crash monologue.


sql/mysqld.cc:
  On a crash when --memlock was active, emit advice about the insta-
  bility of that parameter.

into  rolltop.ignatz42.dyndns.org:/mnt/storeage/mysql-4.1-maint


mysql-test/mysql-test-run.pl:
  Auto merged

Test Suite failing on Windows because the $opt_report_features variable was not defined before use. 


mysql-test/mysql-test-run.pl:
  Defined global variable.

Improve log message

into  rolltop.ignatz42.dyndns.org:/mnt/storeage/mysql-4.1-maint


mysql-test/mysql-test-run.pl:
  Auto merged

As get_arg0_date() in the Item_func_last_day::get_date() returns 
0000-00-00 date sometimes, we have to check ltime->month for 0 after the call.


mysql-test/r/func_time.result:
  Fix for bug #23653: Crash if last_day('0000-00-00')
    - test result.
mysql-test/t/func_time.test:
  Fix for bug #23653: Crash if last_day('0000-00-00')
    - test case.
sql/item_timefunc.cc:
  Fix for bug #23653: Crash if last_day('0000-00-00')
    - return error if month is 0.