Forgot to release an x-latch in DROP DATABASE

  Protect InnoDB monitor with a mutex

I wish I had connection stop  / start so that I could have made a 
test case.

  Remove debug code

  Eliminate potential division by zero

  Merge InnoDB-3.23.52b

into sergbook.mysql.com:/usr/home/serg/Abk/mysql

(as when previous myrg_rkey failed, e.g. found nothing)

  Remove a debug printf which unintentionally was committed to the source tree

  Clarify the setting of innodb_data_file_path in my.cnf: it is in the [mysqld] section

  Remove freeing of all memory at shutdown: since we do not exit the i/o-handlers and other threads, we could potentially get a seg fault at shutdown

  The UNIV_DEBUG version asserted if a foreign key check failed because of a lock wait timeout
btr0cur.c:
  Fix a bug: if the primary key was updated (or delete + insert) so that only the case of characters changed, we were left with a dangling adaptive hash index pointer, often causing an assertion failure in page0page.ic line 515

into hundin.mysql.fi:/home/heikki/mysql3

  Fix a bug: if buffer pool size > 2 GB on a 32-bit computer, we got an assertion failure because >> acted as a division operator (not as a register shift) for a wrapped-over integer value

  Crash the server if file flush does not succeed: the database can get corrupt, better to crash it than let it run

  Slave thread is always in an auto-commit state: remove unnecessary tests when we flush master info

  Initing a mutex with MY_MUTEX_INIT_FAST may relieve thread thrashing on Linux

  Merge 3.23.52

  Use current_thd to check if a transaction is done by a slave thread

  Add syntax SET NO_FOREIGN_KEY_CHECKS=1 and SET RELAXED_UNIQUE_CHECKS=1

  Add option innodb_flush_log_at_trx_commit=2 which means we always write the log to file at commit, but only do fsync about once per second

  Merge 3.23.52

  Add BEGIN andd COMMIT around transactions in the binlog

  Do not flush the position to master.info file if we have a transaction open

configure.in

into mysql.com:/my/mysql-3.23

has finally been released 

header in x86 assemler files instead of eating the whole file