which is always false. In a very fast InnoDB shutdown, we just ensure that
no more transactions are running, flush InnoDB log, signal InnoDB threads to die,
and then return from InnoDB (from innobase_end()) without waiting for those threads
to actually die. I have tested on a 4CPU machine that even with --innodb_flush_log_at_trx_commit=0,
this optimized InnoDB very fast shutdown loses no committed transactions. Patch pre-approved by Heikki.

  Add dummy changelog file. It should merge upward without conflict.

into mysql.com:/home/pem/work/mysql-5.0

into mysql.com:/M50/mysql-5.0

into mysql.com:/home/pem/work/mysql-5.0

into ua141d10.elisa.omakaista.fi:/home/my/bk/mysql-5.0

Reverted Jim's patch.

- bumped up version number in configure.in from 4.0.24 -> 4.0.25 (ChangeSet 1.2073 was already tagged
  as 4.0.24)

into mysql.com:/home/pem/work/mysql-5.0

into hundin.mysql.fi:/home/marko/mysql-5.0

(lost in merge because the code to open these logs was moved to a different
file).

into hundin.mysql.fi:/home/marko/mysql-5.0

into gluh.mysql.r18.ru:/home/gluh/MySQL-BUGS/mysql-4.0.7498

into hundin.mysql.fi:/home/marko/mysql-5.0

  row_search_for_mysql(): Add a clarifying comment.

Actually problem was not in AVG function, but in SUM before the AVG in the
query.

(whereas it's 0 in 4.1, for safety as this is new code).

into mysql.com:/home/svoj/devel/mysql/ft-mysql-5.0

  ut_malloc_low(): fix the fprintf() format string
  (an extraneous comma split the format into two arguments)

The patch implements the idea suggested by Olaf van der Spek in 
thread "Client: many small reads?" (internals@lists.mysql.com).
Now small reads performed by the client library are buffered.
The buffering gives up to 2 times speedup when retrieving 
one-column tables.

  Comment changes: use standard style; fix grammar.

  after merge fixes

into mysql.com:/home/bar/mysql-5.0

into mysql.com:/home/dlenev/src/mysql-5.0-sp

Improved handling of situations when we encounter error during
CREATE PROCEDURE (FUNCTION/TRIGGER/...) and bail out of yyparse()
without restoring proper THD::lex.

  Logging to logging@openlogging.org accepted
func_group.result, func_group.test:
  Added a test case for bug #8893.
opt_sum.cc:
  A misplaced initialization for the returned parameter
  prefix_len in the function find_key_for_maxmin caused
  usage of a wrong key prefix by the min/max optimization
  in cases when the matching index was not the first index
  that contained the min/max field.