signal. Tested to the extent that it does not break Linux. Could not test on
Darwin because the machine was not reachable. Will push anyway - at least it
does not make things worse. Testing on Darwin is needed, however.

it uses now thd->lex.ftfunc_list
moved ft-initialization to a separate function
re-disabled implicit ft initialization

  Fix a seg fault if a lock wait timeout is exceeded at the some time someone grants the lock

was up to date.

  Fix wrong (though equivalent) type decl

  Fix a bug in insert buffer B-tree upper levels; probably caused the crash by B.Fitzpatrick
buf0flu.c:
  Fix a bug in previous change
  A small optimization for LRU flushes to avoid losing hot pages from the buffer pool

  Fix a bug which could cause InnoDB to complain it cannot find free blocks from the buffer cache during recovery

  Fix a typo

  Make database and table names always lower case on Windows

Applied patch to mysqlhotcopy

Fixed problem with t1 LEFT_JOIN t2 ... WHERE t2.date_column IS NULL when date_column is declared as NOT NULL.

Fixed bug with BLOB keys in BDB tables
Fixed problem with MERGE tables on OS with 32 bit files
Fixed that TIME_TO_SEC() works with negative times

Fixed Rows_examined in slow query log.

Fix for consistent 0000-00-00 date handling
Close + open binary logs on flush tables
Fix for AVG() in HAVING.

Use LONG_TIMEOUT (one year) instead of ~0 for long timeouts
Fixed error messages.
Fixed problem with const propagation when comparing columns of different types

Fix for using quoted table names with the SJIS character set.

into mysql.sashanet.com:/home/sasha/src/bk/mysql