Move handling of suffix_length from strnxfrm_bin() to filesort to ensure proper sorting of all kind of binary objects
field::sort_key() now adds length last for varbinary/blob
VARBINARY/BLOB is now sorted by filesort so that shorter strings comes before longer ones
Fixed issues in test cases from last merge

into  mysql.com:/home/my/mysql-5.0

into  mysql.com:/home/my/mysql-5.0

into  mysql.com:/usr/home/bar/mysql-5.0.b9278

into  mysql.com:/usr/home/bar/mysql-5.0.b9278

into  ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-5.0

into poseidon.ndb.mysql.com:/home/tomas/mysql-5.0

  after merge fix.
range.result:
  fixing result accordingly
  ,
cast.result:
  after merge fix
range.test:
  Avoid SELECT'ing of BINARY column not to output 0x00 bytes.

into  ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-5.0

into deer.(none):/home/hf/work/mysql-4.1.clean

into  mysql.com:/home/my/mysql-4.0

- No need to check *ref when ref is checked (Simple optimization fix)

- ensure in ndb_mgmd that the stop command is not issued if a node is restarting
- added some new error messages
- in ndbcntr on master check so that node does not shutdown id shutdoen in progress

into  ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-4.1-tmp

into  mysql.com:/usr/home/bar/mysql-5.0.b9278

  new file
mysql_fix_privilege_tables.sql, mysql_create_system_tables.sh:
  Adding true BINARY/VARBINARY: fixing "password" type, not to be 0x00-padding.
Many files:
  Adding true BINARY/VARBINARY: fixing tests not to output 0x00 bytes.
  Adding true BINARY/VARBINARY: new pad_char structure member.
ctype-bin.c:
  Adding true BINARY/VARBINARY: new pad_char structure member.
  New strnxfrm, with two trailing length bytes.
field.cc:
  Adding true BINARY/VARBINARY.

into moonbone.local:/work/13535-bug-4.1-mysql

into moonbone.local:/work/7672-bug-5.0-mysql

into mysql.com:/Users/kent/mysql/bk/mysql-4.1

into  mysql.com:/opt/local/work/mysql-5.0-12736

into mysql.com:/Users/kent/mysql/bk/mysql-5.0

into mysql.com:/Users/kent/mysql/bk/mysql-5.0

Updated more DDL statements to cause implicit commit.

into  mysql.com:/usr/home/bar/mysql-4.1.b12476

into  mysql.com:/opt/local/work/mysql-5.0-12736

into poseidon.ndb.mysql.com:/home/tomas/mysql-5.0

The bug was in JOIN::join_free which was wrongly determining that
all joins have been already executed and therefore all used tables
can be closed.