into moonbone.local:/work/mysql-4.1-bug-11868

disabled if ref is built with a key from the updated table

Problem was in add_not_null_conds() optimization function.
It contains following code:
JOIN_TAB *referred_tab= not_null_item->field->table->reginfo.join_tab;
...
add_cond_and_fix(&referred_tab->select_cond, notnull);
For UPDATE described in bug report referred_tab is 0 and dereferencing it
crashes the server.



sql/sql_select.cc:
  Fix bug #11868 NOT NULL ref optimization in subquery used in update must be
  disabled if ref is built with a key from the updated table
mysql-test/r/update.result:
  Test case for bug#11868 Update with subquery with ref built with a key from
  the updated table crashes server.
mysql-test/t/update.test:
  Test case for bug#11868 Update with subquery with ref built with a key from the updated table crashes server

into  mysql.com:/opt/local/work/mysql-4.1-root

into  linux.site:/home/reggie/bk/bug7403

 of system vars at PREPARE time": implement a special Item
to handle system variables. This item substitutes itself with 
a basic constant containing variable value at fix_fields.


mysql-test/r/ps.result:
  - test results fixed (Bug#9359).
mysql-test/t/ps.test:
  - add a test case for Bug#9359 "Prepared statements take snapshot
   of system vars at PREPARE time"
sql/item_func.cc:
  - implement Item_func_get_system_var: we should not evaluate system
  variables in the parser, but instead should create an item which 
  is evaluated to a constant at execute.
  - remove an unused function
sql/item_func.h:
  Add a new item, Item_func_get_system_var
sql/mysql_priv.h:
  Move necessary declarations to make set_var.h objects visible in 
  item_func.h
sql/set_var.cc:
  - we should not print to network from get_system_var: if it's called
  from prepared statement prepare, we get packets out of order when using
  the binary protocol. Instead report the error to be sent to the user later.
  This is a backport from 5.0.
sql/set_var.h:
  - declaration of enum_var_type moved to mysql_priv.h

  than in previous 4.1.x

Wrongly applied optimization were adding NOT NULL constraint which results in
rejecting valid rows and reduced result set.

The problem was that add_notnull_conds() while checking subquery were adding
NOT NULL constraint to left joined table, to which, normally, optimization 
don't have to be applied.


sql/sql_select.cc:
  Fix bug #11482 Wrongly applied optimization was erroneously rejecting valid rows
  Constraint were added to optimization appliance test.
mysql-test/t/select.test:
  Test case for bug #11482  Wrongly applied optimization was erroneously rejecting valid rows
mysql-test/r/select.result:
  Test case for bug #11482  Wrongly applied optimization was erroneously rejecting valid rows

into  linux.site:/home/reggie/bk/bug7403

into poseidon.ndb.mysql.com:/home/tomas/mysql-4.1-new


tests/mysql_client_test.c:
  Auto merged

into poseidon.ndb.mysql.com:/home/tomas/mysql-4.1-new


mysql-test/t/ctype_utf8.test:
  Auto merged
ndb/src/common/mgmcommon/ConfigRetriever.cpp:
  Auto merged
ndb/src/common/transporter/TransporterRegistry.cpp:
  Auto merged
ndb/src/mgmclient/CommandInterpreter.cpp:
  Auto merged
ndb/src/mgmsrv/MgmtSrvr.cpp:
  Auto merged
ndb/src/mgmsrv/MgmtSrvr.hpp:
  Auto merged
ndb/src/mgmsrv/Services.cpp:
  Auto merged
sql/sql_prepare.cc:
  Auto merged

into poseidon.ndb.mysql.com:/home/tomas/mysql-4.1-release

      + added close of some fd's
      + debug prints

In 4.1 the connection to the management server _must_ stay, that is how we
ensure that the nodeids are reserved correctly

 BUG#11596 (partial fix), IP addresses not shown in ndb_mgm SHOW command on second ndb_mgmd (or on ndb_mgmd restart)

  mysql user account to that group in the server postinstall section of the RPM spec
  file


support-files/mysql.spec.sh:
  - manual merge of the fix for BUG#10984: create a "mysql" user group and assign the 
    mysql user account to that group in the server postinstall section.

- create a "mysql" user group and assign the mysql user account to that group
  in the server postinstall section. (BUG 10984)
- backported some changes from the 4.1 RPM spec file (cleanups: removed the incomplete
  Brazilian translations, more consequent use of macros)


support-files/mysql.spec.sh:
  - create a "mysql" user group and assign the mysql user account to that group
    in the server postinstall section. (BUG 10984)
  - backported some changes from the 4.1 RPM spec file (cleanups: removed the incomplete
    Brazilian translations, more consequent use of macros)

Fix handling of floats and doubles when using prepared statements             
API in the embedded server. 


sql/sql_prepare.cc:
  Jim's fix for the #10443.
  
  Within the embedded server, there's no need to use float4get()              
  and float8get() for setting parameters, since they are never                
  stored.

mysql-test/t/ps.test:
  Disable prepared statements for a part of the test to avoid
  result mismatch.

BitKeeper/deleted/.del-var~507a9bd3b926223e:
  Delete: mysql-test/var

into  mysql.com:/opt/local/work/mysql-4.1-root

 syntax in binlog which stops replication":
disallow the use of parameter markers which can lead to generation
of malformed binlog queries. 


mysql-test/r/ps.result:
  Test results fixed: a test case for Bug#11299
mysql-test/t/ps.test:
  A test case for Bug#11299
sql/sql_lex.cc:
  Introduce a new parser token for a parameter marker. Make sure
  that a parameter marker can not be used in a query which, when
  transformed to a binlog query, becomes grammatically incorrect.
sql/sql_yacc.yy:
  The check for COM_PREPARE has been moved into the lexer.
mysql-test/var:
  New BitKeeper file ``mysql-test/var''

into mysql.com:/space/my/mysql-4.1.13-clone

    added using ndb_end and showing mem allocation stat at exit
    added init of variable
    added destruction of LocalConfig to still memleak
    added delete of theFacade on exit
    stopSessions(true) on socket server to ensure destuction of threads
    added destruction of arbit manager thread
    added true option to stopSessions in transporter facade to ensure destructions of threads


ndb/include/util/ndb_opts.h:
  added using ndb_end and showing mem allocation stat at exit
ndb/src/common/logger/LogHandler.cpp:
  added init of variable
ndb/src/mgmapi/mgmapi.cpp:
  added destruction of LocalConfig to still memleak
  debug printouts
ndb/src/mgmclient/main.cpp:
  added using ndb_end and showing mem allocation stat at exit
ndb/src/mgmsrv/MgmtSrvr.cpp:
  added delete of theFacade on exit
  removed area51
  added debug prints
ndb/src/mgmsrv/main.cpp:
  added using ndb_end and showing mem allocation stat at exit
  moved glob struct to heap to enable ndb_end usage
  moved options out of glob struct
  stopSessions(true) on socket server to ensure destuction of threads
ndb/src/ndbapi/ClusterMgr.cpp:
  added destruction of arbit manager thread
  debug printouts
ndb/src/ndbapi/TransporterFacade.cpp:
  added true option to stopSessions in transporter facade to ensure destructions of threads
  + debug prints

    + added close of some fd's
    + debug prints


ndb/src/mgmclient/CommandInterpreter.cpp:
   BUG#11898 ndb_mgmd not releasing resources, added "ping"
  + debug prints
ndb/src/mgmsrv/MgmtSrvr.hpp:
   BUG#11898 ndb_mgmd not releasing resources, added "ping"
  + debug prints
ndb/src/mgmsrv/Services.cpp:
   BUG#11898 ndb_mgmd not releasing resources, added "ping" on add_listener
  + added close of some fd's
  + debug prints

into  linux.site:/home/reggie/bk/mysql-4.1

information about error".


libmysql/libmysql.c:
  Clear the last error on the statement if mysql_stmt_reset succeeded.
tests/mysql_client_test.c:
  A test case for Bug#11183 "mysql_stmt_reset() doesn't reset information 
  about error"

    added __LINE__ so systemErrorLab for better error printout


ndb/include/ndb_version.h.in:
  stilled mem leak in usage on getVersionString
ndb/include/util/version.h:
  stilled mem leak in usage on getVersionString
ndb/src/common/util/version.c:
  stilled mem leak in usage on getVersionString
ndb/src/kernel/blocks/qmgr/Qmgr.hpp:
  added __LINE__ so systemErrorLab for better error printout
ndb/src/kernel/blocks/qmgr/QmgrMain.cpp:
  stilled mem leak in usage on getVersionString
  added __LINE__ so systemErrorLab for better error printout
ndb/tools/restore/restore_main.cpp:
  stilled mem leak in usage on getVersionString

ndb/include/portlib/NdbTCP.h:
  added debug prints
ndb/include/util/SocketServer.hpp:
  added debug prints
ndb/src/common/mgmcommon/ConfigRetriever.cpp:
  debug prints
ndb/src/common/mgmcommon/IPCConfig.cpp:
  debug prints
ndb/src/common/portlib/NdbMutex.c:
  debug prints
ndb/src/common/portlib/NdbTCP.cpp:
  debug printout
ndb/src/common/portlib/NdbThread.c:
  debug printout
ndb/src/common/transporter/TransporterRegistry.cpp:
  debug printout
ndb/src/common/util/Parser.cpp:
  debug printout
ndb/src/common/util/SocketClient.cpp:
  debug printout
ndb/src/common/util/SocketServer.cpp:
  debug printout

into  mysql.com:/opt/local/work/mysql-4.1-root

into mysql.com:/home/psergey/mysql-4.1-merge-from-build


sql/item_func.cc:
  Auto merged

  include InnoDB)

tests/mysql_client_test.c:
  fix for #11808 backported.
  cmp_double() introduced in order to avoid using long double registers           
  (for local double vars).

  source table present (this happens for ORDER BY after UNION)


mysql-test/r/fulltext_order_by.result:
  Testcase for BUG#11869
mysql-test/t/fulltext_order_by.test:
  Testcase for BUG#11869

  This is needed because in some cases range optimization is performed twice and added 
  NOT NULL item must have correct const_table_map() value. 


mysql-test/r/select.result:
  Testcase for BUG#11700
mysql-test/t/select.test:
  Testcase for BUG#11700

The problem was that on Windows, the socket was in a blocking state 
trying to read.  Setting killed=1 is not enough to break the 
socket out of it's read loop.   You have to cancel the read request
or close the socket (which close_connection does).


sql/mysqld.cc:
  call close_connection on the thread instead of setting killed=1

into  mysql.com:/opt/local/work/mysql-4.1-root

No separate typecode for MEDIUMTEXT/LONGTEXT is added, as we
have no sound decision yet what typecodes and for what types are
sent by the server (aka what constitutes a distinct type in MySQL).



mysql-test/r/ps_2myisam.result:
  Test results fixed: new longtext/longblob length (2^32)
mysql-test/r/ps_3innodb.result:
  Test results fixed: new longtext/longblob length (2^32)
mysql-test/r/ps_4heap.result:
  Test results fixed: new longtext/longblob length (2^32)
mysql-test/r/ps_5merge.result:
  Test results fixed: new longtext/longblob length (2^32)
mysql-test/r/ps_6bdb.result:
  Test results fixed: new longtext/longblob length (2^32)
mysql-test/r/ps_7ndb.result:
  Test results fixed: new longtext/longblob length (2^32)
sql/field.cc:
  A fix for Bug#9735 "mysql_fetch_fields() acts strange on 
  LONGBLOB/LONGTEXT": fix wrong initialization of field_length
  in case of BLOB fields.
tests/mysql_client_test.c:
  A test case for Bug#9735  "mysql_fetch_fields() acts strange on 
  LONGBLOB/LONGTEXT"

into  mysql.com:/usr/home/bar/mysql-4.1.num-conv


mysql-test/r/ctype_utf8.result:
  Auto merged
mysql-test/t/ctype_utf8.test:
  Auto merged
sql/item.cc:
  Auto merged
sql/item.h:
  Auto merged
sql/sql_table.cc:
  Auto merged

character set is UCS2".
The bug is no longer repeatable.


mysql-test/r/ps.result:
  Test results fixed (a test case for Bug#9442)
mysql-test/t/ps.test:
  A test case for Bug#9442 "Set parameter make query fail if column 
  character set is UCS2