Add policy directive about keeping mysqltest framework tools identical in all versions

  See the "MySQL Test framework manual" for more information

  http://dev.mysql.com/doc/mysqltest/en/index.html

  Please keep the test framework tools identical in all versions!

  Written by:

  Sasha Pachev <sasha@mysql.com>

  Matt Wagner  <matt@mysql.com>

#!/usr/bin/perl

# -*- cperl -*-

# This is a transformation of the "mysql-test-run" Bourne shell script

# to Perl. There are reasons this rewrite is not the prettiest Perl

# you have seen

#

#   - The original script is huge and for most part uncommented,

#     not even a usage description of the flags.

##############################################################################

#

#  mysql-test-run.pl

#

#  Tool used for executing a suite of .test file

#

#   - There has been an attempt to write a replacement in C for the

#     original Bourne shell script. It was kind of working but lacked

#     lot of functionality to really be a replacement. Not to redo

#     that mistake and catch all the obscure features of the original

#     script, the rewrite in Perl is more close to the original script

#     meaning it also share some of the ugly parts as well.

#  See the "MySQL Test framework manual" for more information

#  http://dev.mysql.com/doc/mysqltest/en/index.html

#

#   - The original intention was that this script was to be a prototype

#     to be the base for a new C version with full functionality. Since

#     then it was decided that the Perl version should replace the

#     Bourne shell version, but the Perl style still reflects the wish

#     to make the Perl to C step easy.

#  Please keep the test framework tools identical in all versions!

#

##############################################################################

#

# Some coding style from the original intent has been kept

# Coding style directions for this perl script

#

#   - To make this Perl script easy to alter even for those that not

#     code Perl that often, the coding style is as close as possible to

#     code Perl that often, keeep the coding style as close as possible to

#     the C/C++ MySQL coding standard.

#

#   - All lists of arguments to send to commands are Perl lists/arrays,

#     the information. This separates the "find information" from the

#     "do the work" and makes the program more easy to maintain.

#

#   - At the moment, there are tons of "global" variables that control

#     this script, even accessed from the files in "lib/*.pl". This

#     will change over time, for now global variables are used instead

#     of using %opt, %path and %exe hashes, because I want more

#     compile time checking, that hashes would not give me. Once this

#     script is debugged, hashes will be used and passed as parameters

#     to functions, to more closely mimic how it would be coded in C

#     using structs.

#

#   - The rule when it comes to the logic of this program is

#

#       command_line_setup() - is to handle the logic between flags

  my $tinfo= shift;

  my $tname= $tinfo->{'name'};

  mtr_tofile($path_mysqltest_log,"CURRENT TEST $tname\n");

  # Save info from this testcase run to mysqltest.log

  mtr_appendfile_to_file($path_timefile, $path_mysqltest_log)

    if -f $path_timefile;

  mtr_tofile($path_mysqltest_log,"CURRENT TEST $tname\n");

}