Remove unintentional changes in previous push

  Fix bug: if one used the rename trick of ibman section 15.1 to recover a temp table, InnoDB asserted because it tried to lock the data dictionary twice

into ice.snake.net:/Users/paul/mysql-4.0

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

Rotate_log_event::exec_event() believed that the relay log was corrupted. Fixed it
by moving the test for corruption to Start_log_event::exec_event(). 
Changed Rotate_log_event::exec_event() to not increment positions when the
event is seen in the middle of a transaction.
I did a separate commit in 4.1 (so this should not be merged to 4.0) because
code is a bit different in 4.1.
A test to see if the slave detects when the master died while writing a 
transaction to the binlog (uses a forged truncated binlog I made).

into ice.snake.net:/Volumes/ice2/MySQL/bk/mysql-4.0

into mishka.mysql.fi:/home/my/mysql-4.0

Use forced close of socket to make mysqld shutdown faster when used under valgrind

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

and a comment to explain why.

into eagle.mysql.r18.ru:/home/vva/work/BUG_1194/mysql-4.0

 ::native_val, ::native_val_int and native_val_str into inline

with less code duplication

is involving an assembly function.
Testcase was BUG#1194
SELECT @A:=1; SELECT @A, @A:=1;

(a missing synchronization in the test)

Added resetting of replace_result patterns after --exec command in mysqltest (in order to be consistent with ordinary query execution).

   from ulong to ulonglong by defining -DBIG_JOINS

Add quoting for use `database` for mysqlbinlog
Removed test ins0000001
Add support for --replace for exec in mysqltest
Don't refer to install dir in mysqlbinlog.result

it's just backquoting the db's and table's names
when writing DROP TEMPORARY TABLE to the binlog when a connection ends.
A testcase for this.

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

I am almost sure this is the cause for
BUG#1381 [Opn]: Bug in replication on HP-UX 64 bit binaries?
BUG#1256 [CRp]: Replication slave fails to connect to master in 64-bit version (Solaris)
The reason why I think it's wrong is that the normal client code has 
uint32 ip_addr.
(of course on 32-bit machines it does not matter, but on 64-bit it does).

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

   the calls of external programs "uname" and "pwd" with Perl builtin
   functions "cwd()" and "POSIX::uname()"

into mysql.com:/home/dlenev/src/mysql-4.0-binlog

(changes in Item_func_set_user_var::update, ::val, ::val_str, ::val_int)

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

"If LOAD DATA INFILE 'small_file' fails on master, slave leaves temp files"
(the bug is in the master)

into mysql.com:/home/dlenev/src/mysql-4.0-bgrn

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

Bug #1392 "On Win, slave leaves one temp file after successf. replicating LOAD DATA INFILE"
Windows-specific bug (we forgot to close a file before deleting it).
Patch written by me, tested by Miguel (thanks!) and it works.

Particularly for BUG#1340 mysqlbinlog coredumps when reading Exec
event without seeing Create_file event before