Bug#44653: Server crash noticed when executing random queries with partitions.

When opening a table, it is imperative that the flag
TABLE::auto_increment_field_not_null be false. But if an error occured during
the creation of a table (e.g. the table exists already) with an auto_increment
column and a BEFORE trigger that used the INSERT ... SELECT construct, the
flag was not reset until after error checking. Thus if an error occured,
select_insert::send_data() returned immediately and it was not reset (see * in
pseudocode below).  Crash happened if the table was opened again. Fixed by
resetting the flag after error checking.

nested-loops_join():
  for each row in SELECT table {
    select_insert::send_data():
      if a values is supplied for AUTO_INCREMENT column
         table->auto_increment_field_not_null= TRUE
       else
         table->auto_increment_field_not_null= FALSE
       if (error)
         return 1; *
       if (table->auto_increment_field_not_null == FALSE)
         ...
       table->auto_increment_field_not_null == FALSE 
  }
<-- table returned to table cache and later retrieved by open_table: 
open_table():
  assert(table->auto_increment_field_not_null)


mysql-test/r/trigger.result:
  Bug#44653: Test result
mysql-test/t/trigger.test:
  Bug#44653: Test case
sql/sql_insert.cc:
  Bug#44653: Fix: Make sure to unset this field before returning in case of error

1. BUG#45357 - 5.1.35 crashes with Failing assertion: index->type & DICT_CLUSTERED

2. Also fixes the compilation problem when the flag -DUNIV_MUST_NOT_INLINE

Detailed revision comments:

r5340 | marko | 2009-06-17 12:11:49 +0300 (Wed, 17 Jun 2009) | 4 lines
branches/5.1: row_unlock_for_mysql(): When the clustered index is unknown,
refuse to unlock the record.
(Bug #45357, caused by the fix of Bug #39320).
rb://132 approved by Sunny Bains.
r5339 | marko | 2009-06-17 11:01:37 +0300 (Wed, 17 Jun 2009) | 2 lines
branches/5.1: Add missing #include "mtr0log.h" so that the code compiles
with -DUNIV_MUST_NOT_INLINE.

The merge from http://lists.mysql.com/commits/76678 caused the 
growth_size parameter to the my_init_dynamic_array function to
be ignored. This patch corrects the problem.

mysys/hash.c:
  Bug#43572  Handle failures from hash_init
  
  Replacing the last parameter to my_init_dynamic_array with
  growth_size.

  Details:
  - Limit the queries to character sets and collations
    which are most probably available in all build types.
    But try to preserve the intention of the tests.
  - Remove the variants adjusted to some build types.

  Note:
  1. The results of the review by Bar are included.
  2. I am not able to check the correctness of this patch
     on any existing build type and any MySQL version.
     So it could happen that the new test fails somewhere.

occasionally.

mysql_multi can call mysqld_safe. In doing this it's not changing the 
current working directory. This may cause confusion in the case where 
mysqld_multi is handling instances of servers of different versions 
and the current working directory is the installation directory of one 
of these servers.

Fixed by enhancing the meaning of basedir in [mysqldN] sections of 
mysqld_multi. If specified, mysqld_multi will change the current 
working directory to the basedir directory before starting the server 
in mysqld_multi ... start ... and then change it back to what it was.

scripts/mysqld_multi.sh:
  Bug #36654: optionally preserve, change and restore the cwd when 
  starting server instances

      
Failure to allocate memory for the hash->array element,
caused hash_init to return without initializing the other
members of the hash. Thus although the dynamic array
buffer may be allocated at a later point in the code, the
incompletely initialized hash caused fatal failures.

This patch moves the initialization of the other members
of the hash above the array allocation, so that the usage
of this hash will not result in fatal failures.

include/hash.h:
  Bug#43572 Handle failures from hash_init
  
  hash_inited is used to verify that the hash is
  valid. After the change induced by the current
  patch hash->array.buffer !=0 is not a valid check
  for this condition, since, the dynamic array can
  be allocated even at a later time. Bootstrap SQL
  script is setting some variables, which are
  actually not set due to this hash_inited issue.
  Thus we get empty grant tables.
  
  A better way to check if the hash is valid is
  to verify that hash->blength is greater than 0.
mysys/hash.c:
  Bug#43572 Handle failures from hash_init
  
  Move the initialization of the other members
  of the hash above the array allocation, so that
  the usage of this hash will not result in fatal
  failures.

      
Inconsistent behavior of session variable max_allowed_packet 
(and net_buffer_length); only assignment to the global variable 
has any effect, without this being obvious to the user.
      
The patch for Bug#22891 is backported to 5.0, making the two
session variables read-only. As this is a backport to GA 
software, the error used when trying to assign to the read-
only variable is ER_UNKNOWN_ERROR. The error message is the 
same as in 5.1+.

mysql-test/t/variables.test:
  Tests are changed to account for the new semantics, and assignment to the read-only variables is added to test 
  the emission of the correct error message.
sql/set_var.cc:
  Both max_allowed_packet and net_buffer_length are changed 
  to be of type sys_var_thd_ulong_session_readonly. ER_UNKNOWN_ERROR is used to indicate an attempt to assign 
  to an instance of a read-only variable.
sql/set_var.h:
  Class sys_var_thd_ulong_session_readonly is added.

The problem: described in the bug report.
The fix:
--increase buffers where it's necessary
  (buffers which are used in stxnmov)
--decrease buffer lengths which are used


client/mysql.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/ha_ndbcluster.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/ha_ndbcluster_binlog.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/handler.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/log.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/mysqld.cc:
  removed unnecessary line
sql/parse_file.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_acl.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_base.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_db.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_delete.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_partition.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_rename.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_show.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_table.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function
sql/sql_view.cc:
  --increase buffers where it's necessary
    (buffers which are used in stxnmov)
  --decrease buffer lengths which are used
    as argument for strxnmov function

The problem was that a pthread.h header used by gcc did not
declare the pthread_setscheprio, yet the function is implemented
by the function is implemented, causing a autoconf check to pass
and compilation with C++ to fail. The solution is to add a
autoconf check to ensure that the function is properly declared.

configure.in:
  Check that the pthread_setschedprio is declared.

Large transactions and statements may corrupt the binary log if the size of the
cache, which is set by the max_binlog_cache_size, is not enough to store the
the changes.

In a nutshell, to fix the bug, we save the position of the next character in the
cache before starting processing a statement. If there is a problem, we simply
restore the position thus removing any effect of the statement from the cache.
Unfortunately, to avoid corrupting the binary log, we may end up loosing changes
on non-transactional tables if they do not fit in the cache. In such cases, we
store an Incident_log_event in order to stop the slave and alert users that some
changes were not logged.

Precisely, for every non-transactional changes that do not fit into the cache,
we do the following:
  a) the statement is *not* logged
  b) an incident event is logged after committing/rolling back the transaction,
  if any. Note that if a failure happens before writing the incident event to
  the binary log, the slave will not stop and the master will not have reported
  any error.
  c) its respective statement gives an error

For transactional changes that do not fit into the cache, we do the following:
  a) the statement is *not* logged
  b) its respective statement gives an error

To work properly, this patch requires two additional things. Firstly, callers to
MYSQL_BIN_LOG::write and THD::binlog_query must handle any error returned and
take the appropriate actions such as undoing the effects of a statement. We
already changed some calls in the sql_insert.cc, sql_update.cc and sql_insert.cc
modules but the remaining calls spread all over the code should be handled in
BUG#37148. Secondly, statements must be either classified as DDL or DML because
DDLs that do not get into the cache must generate an incident event since they
cannot be rolled back.