When loading dump created by mysqldump tool an error is
thrown saying storage engine for the table doesn't have 
an option.
                        
mysqldump tries to re-insert the data into the federated
table which causes the error. Since the data is already
available on the remote server, mysqldump shouldn't try
to dump the data again for FEDERATED tables.
                        
As stated in the bug page, it can be considered similar
to the MERGE ENGINE with "view only" nature.
                        
Fixed by adding the "FEDERATED ENGINE" to the exception
list to ignore the data.

client/mysqldump.c:
  Fixed check_if_ignore_table() to ignore FEDERATED engine
  when dumping the table data.
mysql-test/r/federated.result:
  Result file for BUG#21360
mysql-test/t/federated.test:
  Testcase for BUG#21360

                   ~40Mb after mysqldump/import 
        
When the input string exceeds the maximum allowed size for the 
internal buffer, batch_readline() returns a truncated string. 
Since there was no way for a caller to determine whether the 
string was truncated or not, the command line client assumed 
batch_readline() to always return the whole input string and 
appended a newline character. This resulted in garbled data 
when importing dumps containing strings longer than the 
maximum input buffer size. 
  
Fixed by adding a flag to the batch_readline() interface to 
signal a truncated string to the caller. 
  
Other minor problems fixed during patch implementation: 
 
- The maximum allowed buffer size for batch_readline() was set 
up depending on the client's max_allowed_packet value. It does 
not actully make any sense, as those variables are not 
related. The input buffer size limit is now always set to 1 
MB. 
  
- fill_buffer() did not always set the EOF flag. 
 
- The input buffer could actually grow twice as the specified 
limit due to insufficient checks in intern_read_line(). 

client/my_readline.h:
  Changed the interface of batch_readline().
client/mysql.cc:
  Honor the truncated flag returned by batch_readline() and do  
  not append the newline character if it was set. Since we can't 
  change the interfaces for readline()/fgets() used in the  
  interactive mode, always assume the returned string was not  
  truncated. In addition, always set the batch_readline()  
  internal buffer to 1 MB, independently from the client's  
  max_allowed_packet.
client/readline.cc:
  Added the 'truncated' argument do batch_readline() to signal 
  truncated string to a caller. 
  Fixed fill_buffer() to set the EOF flag correctly. 
  Fixed checks in intern_read_line() to not allow the internal  
  buffer grow past the specified limit.
mysql-test/r/mysql.result:
  Added a test case for bug #41486.
mysql-test/t/mysql.test:
  Added a test case for bug #41486.

mysql.procs_priv table itself does not get replicated.
Inserting routine privilege record into mysql.procs_priv table
is triggered by creating function/procedure statements
according to current user's privileges.
Because the current user of SQL thread has GLOBAL_ACL,
which doesn't need any check mysql.procs_priv privilege
when create/alter/execute routines.
Corresponding GLOBAL_ACL privilege user
doesn't insert routine privilege record into
mysql.procs_priv when creating a routine.

Fixed by switching the current user of SQL thread to definer user if
the definer user exists on slave.
That populates procs_priv, otherwise to keep the SQL thread
user and procs_priv remains unchanged.


mysql-test/suite/rpl/r/rpl_do_grant.result:
  Test case result for routine privilege when definer user exist or not on slave
mysql-test/suite/rpl/t/rpl_do_grant.test:
  Test case result for routine privilege when definer user exist or not on slave
sql/sql_parse.cc:
  Switch current user of SQL thread to definer user if the definer user
  existes on slave when checking whether the routine privilege is
  needed to insert mysql.procs_priv table or not.

Any statement reading corrupt archive data file
(CHECK/REPAIR/SELECT/UPDATE/DELETE) may cause assertion
failure in debug builds. This assertion has been removed
and an error is returned instead.

Also fixed that CHECK/REPAIR returns vague error message
when it mets corruption in archive data file. This is
fixed by returning proper error code.

mysql-test/r/archive.result:
  A test case for BUG#32880
mysql-test/std_data/bug32880.ARN:
  corrupted archive table to test check and repair table operation
mysql-test/std_data/bug32880.ARZ:
  corrupted archive table to test check and repair table operation
mysql-test/std_data/bug32880.frm:
  corrupted archive table to test check and repair table operation
mysql-test/t/archive.test:
  A test case for BUG#32880
storage/archive/ha_archive.cc:
  Fixed unpack_row() to return the error instead of throwing assertion 
  and also fixed repair() to throw better error when repair table
  operation fails on corrupted archive table

The reason for the error is incorrectly specified link dependencies
for mysql_embedded, mysqltest_embedded and mysql_client_test_embedded
in CMakeLists.txt (ADD_DEPENDENCIES should be TARGET_LINK_LIBRARIES)


libmysqld/CMakeLists.txt:
  changed library type for libmysqld to SHARED instead of
  MODULE. MODULE in CMake notation is a shared library that 
  is used only in dlopen/dlsym/LoadLibrary scenarios.
  Hence it was impossible to use TARGET_LINK_LIBRARIES with
  a MODULE.
libmysqld/examples/CMakeLists.txt:
  Use TARGET_LINK_LIBRARIES (instead of previously incorrectly 
  used ADD_DEPENDENCIES) to specify link dependency from libmysqld

	patch enclosed)
  
One call to my_error_unregister_all() would free pointers, but leave one
pointer to just-freed memory still assigned.  That's the bug.  Subsequent
calls of this function would try to follow pointers into deallocated, 
garbage memory and almost certainly SEGV.

Now, after freeing a linked list, unset the initial pointer.

become negative

- merged the fix to 5.1
- extended to cover I_S.PROCESSLIST.TIME
- Changed the column type of I_S.PROCESSLIST.TIME from LOGNLONG
  UNSIGNED
  to LONG (to match the SHOW PROCESSLIST type)
- Added a test case

"Release_lock("hello")" is now also successful when delivering NULL, replaced two sleeps by wait_condition. The last two "sleep 1" have not been replaced as all tried wait conditions leaded to nondeterministic results, especially to succeeding concurrent updates. To replace the sleeps there should be some time planned (or internal knowledge of the server may help).
 

sql/sql_select.cc:
  the if() separate arguments checks are slightly faster.

Re-recording of .result files for additional tests that use this include file.