This fixed some optimization problems when using -DBIG_TABLES
Portabilty fixes for OpenUnix and HPUX
Added C and C++ version numbers to mysqlbug


Docs/mysqld_error.txt:
  Added new error message
acinclude.m4:
  Fix for configure problem on OpenUnix
configure.in:
  Fix for OpenUnix
  Added C and C++ versions to mysqlbug
mysql-test/r/variables.result:
  Update of max_join_size handling
mysql-test/t/variables.test:
  Update of max_join_size handling
mysys/Makefile.am:
  Removed duplicate row
mysys/my_alloc.c:
  Safety fixes (not fatal)
scripts/Makefile.am:
  Added C and C++ compiler versions to mysqlbug
scripts/mysqlbug.sh:
  Added C and C++ compiler versions to mysqlbug
sql/item_func.cc:
  Fixed that user variables that changes are not threated as constants.
sql/item_func.h:
  Fixed that user variables that changes are not threated as constants.
sql/mysqld.cc:
  Changed thd variables max_join_size and select_limit to type ha_rows
sql/repl_failsafe.cc:
  Removed not needed cast
sql/set_var.cc:
  Changed thd variables max_join_size and select_limit to type ha_rows
sql/set_var.h:
  Changed thd variables max_join_size and select_limit to type ha_rows
sql/slave.cc:
  Removed not needed cast
sql/sql_class.h:
  Changed thd variables max_join_size and select_limit to type ha_rows
sql/sql_parse.cc:
  Removed not needed cast
  Fixed security problem with mysql_drop_db()
sql/sql_show.cc:
  Changed thd variables max_join_size and select_limit to type ha_rows
sql/structs.h:
  Changed thd variables max_join_size and select_limit to type ha_rows

into mysql.com:/home/walrus/bk/40

sql-bench/crash-me.sh:
  add tagline for emacs 
  add log lines to result file
  remove wrong prompt for WEEK()
sql-bench/server-cfg.sh:
  cleanup field conversation

(Accidently disabled by a recent push)

  Add to perror the code 150 about wrong foreign key def


extra/perror.c:
  Add to perror the code 150 about wrong foreign key def

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


mysys/my_rename.c:
  Auto merged

This fixes a bug with REPAIR TABLE t1 USE_FRM on windows.


mysys/my_rename.c:
  On windows, delete any conflicting 'to' files on rename(from,to)

configure.in:
  merge with 3.23
sql/log.cc:
  Use local file
sql/mysqld.cc:
  Get patch for lower_case_table_names for MacOSX

Don't enable any bulk insert or record caching code if inserting less than MIN_ROWS_TO_USE_BULK_INSERT rows (100)


myisam/mi_check.c:
  Fixed bug in copying statistics for disabled index
mysql-test/r/distinct.result:
  Fix result after not doing key statistics for first insert.
mysql-test/r/fulltext.result:
  Fix result after not doing key statistics for first insert.
mysql-test/r/insert.result:
  Fix result after not doing key statistics for first insert.
mysql-test/r/key_diff.result:
  Fix result after not doing key statistics for first insert.
mysql-test/r/order_by.result:
  Fix result after not doing key statistics for first insert.
mysql-test/r/select.result:
  Fix result after not doing key statistics for first insert.
mysql-test/r/show_check.result:
  Fix result after not doing key statistics for first insert.
sql/ha_myisam.cc:
  Don't disable index when inserting only a few rows
sql/log.cc:
  Transactions in AUTOCOMMIT=0 mode didn't rotate binary log.
sql/sql_insert.cc:
  Don't enable any bulk insert or record caching code if inserting less than MIN_ROWS_TO_USE_BULK_INSERT

sql/sql_yacc.yy:
  Fixed syntax error (with bison 1.75).

   in sql/ha_myisammrg.h (removed "ha_myisammrg::")

configure.in:
  merge with 4.0
scripts/mysqld_safe.sh:
  Don't apply patch from 3.23

tables.  This removes a possible deadlock situation.

This fixes a problem in replication where RESET SLAVE could crash a running slave.


mysql-test/r/rpl_failsafe.result:
  Added missing drop
mysql-test/r/rpl_log_pos.result:
  Added slave stop before change master
mysql-test/t/rpl_log_pos.test:
  Added slave stop before change master
sql/sql_parse.cc:
  reset_slave() sends it's own errors
sql/sql_repl.cc:
  reset_slave() sends it own error.
  RESET SLAVE and CHANGE MASTER will now give an error if slave is running.
  This fixes a problem in replication where RESET SLAVE could crash a running slave.

   spotting and reporting this)

Fixed wrong variable name for Windows


VC++Files/innobase/innobase.dsp:
  Don't use precompiled headers (doesn't work with samba)
VC++Files/myisammrg/myisammrg.dsp:
  Add missing files
VC++Files/mysqlbinlog/mysqlbinlog.dsp:
  Fixed included paths
VC++Files/mysqldemb/mysqldemb.dsp:
  Don't use precompiled headers (doesn't work with samba)
VC++Files/mysqlserver/mysqlserver.dsp:
  Don't use precompiled headers (doesn't work with samba)
VC++Files/sql/mysqld.dsp:
  Fixed paths
innobase/include/univ.i:
  Removed compiler warning
myisammrg/myrg_open.c:
  Removed compiler warning
sql/ha_myisam.cc:
  Removed compiler warning
  Fixed wrong (but not dangerous) if statement
sql/ha_myisammrg.cc:
  Removed compiler warning
sql/mysqld.cc:
  Fixed wrong variable name for Windows
sql/sql_update.cc:
  Removed not used label

 - Replace function 'check_reserved_words' with correct one


BitKeeper/etc/logging_ok:
  Logging to logging@openlogging.org accepted

  We must not commit inside InnoDB when LOAD DATA INFILE just writes a block to the binlog; only commit if the log event type is QUERY_EVENT or EXECUTE_LOAD_EVENT


sql/log.cc:
  We must not commit inside InnoDB when LOAD DATA INFILE just writes a block to the binlog; only commit if the log event type is QUERY_EVENT or EXECUTE_LOAD_EVENT

sql/field.cc:
  Small safety fix

 - bumped up version number in configure.in to 3.23.55

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

Added syntax support for
CREATE TABLE foo (a char CHARACTER SET latin1) CHARSET=latin1;


Docs/internals.texi:
  Update binary protocol description
innobase/include/db0err.h:
  Merge from 3.23
mysql-test/r/insert.result:
  Updated test result from 3.23
sql/log.cc:
  Fixed bug in replication and log rotation
sql/log_event.cc:
  Ensure that BEGIN / COMMIT is handled properly if slave dies
sql/slave.cc:
  Fixed bug in replication and log rotation
sql/slave.h:
  Ensure that BEGIN / COMMIT is handled properly if slave dies
sql/sql_analyse.cc:
  Moved usage of res before res is destroyed (by bzero(&s...))
sql/sql_yacc.yy:
  Added syntax support for
  
  CREATE TABLE foo (a char CHARACTER SET latin1) CHARSET=latin1;
  
  To be able to read MySQL 4.1 dump files.

innobase/dict/dict0dict.c:
  Auto merged
libmysql/libmysql.c:
  Auto merged
scripts/make_binary_distribution.sh:
  Auto merged
sql/mysqld.cc:
  Auto merged
sql/sql_udf.cc:
  Auto merged
sql/ha_innodb.cc:
  Auto merged
innobase/include/db0err.h:
  Skiped merge (Heikki wants to do this)
innobase/row/row0mysql.c:
  Skiped merge (Heikki wants to do this)
sql/sql_show.cc:
  Use local code

libmysql/libmysql.c:
  Compatibility fix
sql/mysqld.cc:
  Added back -max if we are using InnoDB.
  (This should not be done in the 3.23 version)

sql/sql_show.cc:
  Fixed rare bug in show processlist which could cause core dump if a new thread was connecting.
  (We where calling mutex_unlock() without having ever called mutex_lock)
sql/sql_udf.cc:
  Fixed warning from valgrind

into serg.mysql.com:/usr/home/serg/Abk/mysql-4.0

  updated


mysql-test/r/key_diff.result:
  updated
mysql-test/r/order_by.result:
  updated

into mysql.com:/home/walrus/bk/40.ibase


BitKeeper/etc/logging_ok:
  auto-union
sql-bench/server-cfg.sh:
  Auto merged