while slave thread has temp tables - not hard to fix though, but it is time to go to bed

added a new test case for temp table replication and Slave_open_temp_tables
status variable.

Misc fixes


client/mysqltest.c:
  added disconnect command
mysql-test/create-test-result:
  fixed bugs so it actually works
mysql-test/t/include/master-slave.inc:
  establish alternate connections to master and slave
sql/mysql_priv.h:
  Slave_open_temp_tables
sql/mysqlbinlog.cc:
  fixed a bug in reading local binlogs
sql/mysqld.cc:
  added Slave_open_temp_tables status variable
sql/slave.cc:
  fixed freed memory reference in the error message
sql/sql_class.cc:
  slave_proxy_id
sql/sql_class.h:
  slave_proxy_id
sql/sql_parse.cc:
  small optimization - reorder eval in &&
BitKeeper/etc/ignore:
  Added mysql-test/r/3.23/rpl000001.b.result.reject mysql-test/r/3.23/rpl000012.result.reject to the ignore list
sql/sql_table.cc:
  added THD arg to rename_temporary_table
sql/unireg.h:
  the table_cache_key is now 4 bytes longer