His explanation:

 The socket on which MySQL listens for new connections on a blocking
  socket most of the time but is set to non-blocking during the
  accept() of the new connection.  Due to a bug in the kernel, the new
  socket returned by accept() is a blocking socket but returns the
  O_NONBLOCK flag when queried via fcntl(F_GETFL).  That is, the file
  descriptor and the underlying socket don't agree on the blocking
  mode.

  Since MySQL determines via fcntl(F_GETFL) that the socket is
  non-blocking, it expects the first read() in my_real_read to not
  block, so it doesn't enable the timeout alarm.  However, the read
  does block, and thus there's no timeout alarm.  The thread kill
  (which relies on rescheduling the timeout alarm) also does not work
  as a consequence.

The bug shows itself if you build MySQL with LinuxThreads support
(needed for SMP on FreeBSD).  Issuing a KILL command in MySQL won't be
"noticed" by the "killed" thread until it runs another query--that
makes KILL pretty useless.  And the wait_timeout doesn't work either.

into sinisa.nasamreza.org:/mnt/work/mysql-4.0

  Flush stderr if we run out of memory, so that the error message more probably finds its way to the error log

into sanja.is.com.ua:/home/bell/mysql/mysql-4.0

excluded double call of 'invalidate()'

 - actually follow the hints in the comment above and "Remember that 
   regexps needs to quote [ and ] since this is run through m4"...

 - fix MYSQL_NO_DASH_VERSION if version number ends on one digit only
   ("comment" test failed when MySQL version was changed from 4.0.5
   to 4.0.5a)

delete from table where column<=>NULL 
on indexed columns

Fix for shutdown on Mac OS X

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

  Fix potential bus error in 64-bit computers in SHOW VARIABLES, if int is only 32-bit in them

  Fix crash when InnoDB temp table is truncated + fix bug: do not X-lock rows under LOCK TABLES except if the table is temp

into genie.(none):/home/vva/work/mysqldump_xml/4.0

into geopard.(none):/storage/bk/my-new-4.0

Redo of changeset I commited, but failed to push

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

 - bumped up version number in configure.in
 - fixed a typo in test-create.sh

into sanja.is.com.ua:/home/bell/mysql/mysql-4.0

 - "make distcheck" needs to be fixed

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

Add support for LIMIT # OFFSET #
Changed lock handling:  Now all locks should be stored in TABLE_LIST instead of passed to functions.
Don't call query_cache_invalidate() twice in some cases
mysql_change_user() now clears states to be equal to close + connect.
Fixed a bug with multi-table-update and multi-table-delete when used with LOCK TABLES
Fixed a bug with replicate-do and UPDATE

  Print an error message if someone tries to call InnoDB startip more than once during process lifetime