fixed bug in load data infile replication

updated test suite to always test if slave obeys stop/start commands 


mysql-test/include/master-slave.inc:
  test if the slave actually stops and starts in response to slave start/slave stop
sql/log_event.cc:
  call the right constructor - I am 99% sure this is what cause sporadic failures of rpl000001
sql/share/romanian/errmsg.sys:
  not changed

mysql-test/include/master-slave.inc

@@ -4,8 +4,10 @@ connect (slave,localhost,root,,test,0,var/tmp/mysql-slave.sock);
 connect (slave1,localhost,root,,test,0,var/tmp/mysql-slave.sock);
 connection slave;
 !slave stop;
+@r/slave-stopped.result show status like 'Slave_running';
 connection master;
 reset master;
 connection slave;
 reset slave;
-!slave start;
+slave start;
+@r/slave-running.result show status like 'Slave_running';

mysql-test/r/slave-running.result

+Variable_name	Value	
+Slave_running	ON	

mysql-test/r/slave-stopped.result

+Variable_name	Value	
+Slave_running	OFF	

sql/log_event.cc

@@ -572,7 +572,7 @@ Load_log_event::Load_log_event(IO_CACHE* file, time_t when, uint32 server_id):
 }
 
 Load_log_event::Load_log_event(const char* buf, int event_len):
-  Log_event(when,0,0,server_id),data_buf(0),num_fields(0),fields(0),
+  Log_event(buf),data_buf(0),num_fields(0),fields(0),
   field_lens(0),field_block_len(0),
   table_name(0),db(0),fname(0)
 {