as the already-stored timestamp. Now 'created' is used only to know if
this is a first binlog or not. And we may re-use the superfluous bytes
in 5.0 when we need room.


sql/log_event.cc:
  This sort of reverts a change I made in 3.23.57. In 3.23.57 I set 'created' to 0
  if this was a non-first binlog, so I made mysqlbinlog not print the "created xx"
  part if created == 0. While this was sensible, as 'created' is 0 or equal to
  the timestamp which is already stored in the event, we can always print
  "created xx" by using the timestamp, and leaving the use of 'created' only to
  know if this is a first binlog or not (which we print as "created at startup").
sql/log_event.h:
  In Start_log_event, 'created' is always the same as 'when', or O.
  We didn't need 4 bytes for this, a bit would have been enough (O or
  "same as 'when'"). Possibly in 5.0 we will re-use the useless bytes.

Otherwise, we'll get questions from users about this curious 1970.

Fix for bug 254 : the first Start_log_event after server startup will
have created=now(), whereas the next ones (FLUSH LOGS, auto rotation)
will have created=0. Before this, it was always now().
This way, slaves >=4.0.14 will know when they must
drop stale temp tables or not. The next task is now modify 4.0.14 to
implement this.


sql/log.cc:
  Fix for bug 254 : the first Start_log_event after server startup will
  have created=now(), whereas the next ones (FLUSH LOGS, auto rotation)
  will have created=0. Before this, it was always now().
  This way, slaves >=4.0.14 will know when they must
  drop stale temp tables or not.
sql/log_event.h:
  An explanation.
sql/sql_class.h:
  Prototype change (see log.cc).

into narttu.mysql.fi:/my/mysql-3.23

(Casused problems in GIS functions in 4.1)


include/global.h:
  Fixed compiler optimization problem with doubleget()

  Prevent the InnoDB main thread from hogging CPU if a table lingers in the background drop queue (though it is essentially a bug if a table end up there at all)


innobase/srv/srv0srv.c:
  Prevent the InnoDB main thread from hogging CPU if a table lingers in the background drop queue (though it is essentially a bug if a table end up there at all)

BitKeeper/deleted/.del-internals.texi~62b6f580a41c2a43:
  Delete: Docs/internals.texi

into narttu.mysql.fi:/my/mysql-3.23

into mashka.mysql.fi:/home/my/mysql-3.23

(this bug was already fixed in 4.0, I just copied and pasted two lines).

into eagle.mysql.r18.ru:/home/vva/work/LOADDATA/mysql-3.23


BitKeeper/etc/logging_ok:
  auto-union
sql/log_event.cc:
  Auto merged

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

into narttu.mysql.fi:/my/mysql-3.23

  Backport from 4.0: We had forgotten to call pthread_mutex_destroy when we free an OS mutex in Unix


innobase/os/os0sync.c:
  Backport from 4.0: We had forgotten to call pthread_mutex_destroy when we free an OS mutex in Unix

sql/sql_show.cc:
  SHOW TABLE STATUS displayed wrong Row_format for myisampack'ed tables.

Security patch to remove wrong error when one had a global update/delete privilige and a database specific SELECT privilege. 


sql/sql_acl.cc:
  Security patch
sql/sql_base.cc:
  Security patch
sql/sql_parse.cc:
  Security patch
tests/grant.pl:
  Test of security patch
tests/grant.res:
  Test of security patch

into narttu.mysql.fi:/my/mysql-3.23

All tests plus grant.pl test passed.

into sinisa.nasamreza.org:/mnt/work/mysql

into serg.mylan:/usr/home/serg/Abk/mysql


client/mysqldump.c:
  Auto merged
BitKeeper/etc/logging_ok:
  Logging to logging@openlogging.org accepted

sql/mysqld.cc:
  Safety fix.

Backported fix from 4.1 for bug 212: SELECT query containing a NATURAL JOIN and parentheses in the WHERE clause


mysql-test/r/join.result:
  New test results
mysql-test/t/join.test:
  Test for bug 212
sql/log_event.cc:
  Removed wrong cast
sql/log_event.h:
  Fixed problem with timestamps in binary log on 64 bit machines
sql/sql_list.h:
  Fix for bug 212 (back ported from 4.1)

into mashka.mysql.fi:/home/my/mysql-3.23


sql/slave.cc:
  Auto merged

Fixed problem with not freed thr_alarm() on slave connect


mysql-test/t/rpl_alter.test:
  Changed name of created test database
sql/ha_innobase.cc:
  Changed innobase_flush_log_at_trx_commit to be 1 as default
sql/mini_client.cc:
  Fixed problem with not freed thr_alarm()
sql/mysqld.cc:
  Allocate some extra thr_alarm slots just to be safe
sql/slave.cc:
  Simple code cleanup

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

mysys/my_vsnprintf.c:
  safety fix
sql/log.cc:
  safety fix

sql/log.cc:
  Portability fix (for hammer)

logged, but read as LOAD DATA INFILE REPLACE"
This was just bad && instead of &, but nasty consequences.
This should be merged to 4.0 BUT it will not be automatic (some code
has moved from log_event.h to log_event.cc, and log_event.cc has changed);
please Merging Man, do 'bk grep REPLACE_FLAG' in 4.0/sql
to find all the new places.


sql/log_event.cc:
  & instead of && when testing flags
sql/slave.cc:
  & instead of && when testing flags

least one BLOB column.

into narttu.mysql.fi:/my/mysql-3.23