2007.04.12

MySQL

MySQL5.1.17ベータ版がリリースされました。

オリジナル版:http://dev.mysql.com/doc/refman/5.1/en/news-5-1-17.html

最も普及しているオープンソースデータベースの新ベータバージョンであるMySQL 5.1.17ベータ版がリリースされました。


これはベータ版であり、他の多くの試作商品リリースと同様、商品レベルのシステムまたは、重要なデータを含むシステムに対してインストールしないように注意してください。5.0使用の商品レベルのシステムについては、以下のMySQL Enterpriseの商品説明のページをご覧ください。

http://mysql.com/products/enterprise/

 

MySQL 5.1.17ベータ版は、http://dev.mysql.com/downloads/とミラーサイトのダウンロード・ページから、ソースコード及び多くのプラットフォームのためのバイナリで現在利用可能です。

すべてのミラーサイトが現在、最新であるとは限らないことに注意してください。
あるミラーサイトでこのバージョンを見つけることができない場合は、再度確認を行うか、あるいは別のダウンロード・サイトを選択してください。

バグレポート、バグ修正、パッチ等の情報をお待ちしておりますので、以下のページをご利用ください。

http://forge.mysql.com/wiki/Contributing

以下のセクションはMySQL5.1の最も最近にリリースされたバージョンであるMySQL5.1.16ベータ版と比較してのMySQLソースコードにおけるバージョン間の変更を記載しています。

http://dev.mysql.com/doc/refman/5.1/en/news-5-1-17.html

以下は、追加または変更された機能です。


Functionality added or changed:

 * Incompatible change: Scheduled events now use the MySQL server
   time zone to determine their schedules, rather than UTC as in
   previous releases. Because of this change, scheduled event
   metadata now includes time zone information, which can be seen
   in the TIME_ZONE column of the INFORMATION_SCHEMA.EVENTS table
   and the Time zone column in the output of the SHOW EVENTS
   statement. These columns have been added in this release,
   along with a time_zone column in the mysql.event table. Due to
   these changes, events created in previous versions of MySQL
   cannot be created, viewed, or used until mysql.event has been
   upgraded. (Bug#16420:http://bugs.mysql.com/16420)
 * Important change: The CREATE EVENT and ALTER EVENT statements
   now support a DEFINER clause, similar to that used in the
   CREATE TRIGGER statement.
   (Bug#16425:http://bugs.mysql.com/16425)
   See Section 20.2.1, "CREATE EVENT Syntax," for detailed
   information.
 * Important change: The following options for controlling
   replication master configuration on a slave are now
   deprecated.
      + --master-host
      + --master-user
      + --master-password
      + --master-port
      + --master-connect-retry
      + --master-ssl
      + --master-ssl-ca
      + --master-ssl-capath
      + --master-ssl-cert
      + --master-ssl-cipher
      + --master-ssl-key
   To change the master configuration on a slave you should use
   the CHANGE MASTER statement. (See also #21490)
 * Statements that affect mysql database tables now are written
   to the binary log using the following rules:
      + Data manipulation statements such as INSERT that change
        data in mysql database tables directly are logged
        according to the settings of the binlog_format system
        variable.
      + Statements such as GRANT that change the mysql database
        indirectly are logged as statements regardless of the
        value of binlog_format.
   For more details, see Section 6.1.2.4, "Logging Format for
   Changes to mysql Database Tables."
   (Bug#25091:http://bugs.mysql.com/25091)
 * Prepared statements now use the query cache under the
   conditions described in Section 5.13.1, "How the Query Cache
   Operates." (Bug#735:http://bugs.mysql.com/735)
 * Added the --secure-file-priv option for mysql-test-run.pl,
   which limits the effect of the load_file command for mysqltest
   and for the LOAD DATA and SELECT ... INTO OUTFILE statements
   to work with files in a given directory.
   (Bug#18628:http://bugs.mysql.com/18628)
 * Added the innodb_stats_on_metadata system variable to enable
   control over whether InnoDB performs statistics gathering when
   metadata statements are executed. See Section 14.5.4, "InnoDB
   Startup Options and System Variables."
   (Bug#26598:http://bugs.mysql.com/26598)
 * Added the hostname system variable, which the server sets at
   startup to the server hostname.
 * The syntax for index hints has been extended to enable more
   fine-grained control over the optimizer's selection of an
   execution plan for various phases of query processing. See
   Section 13.2.7.2, "Index Hint Syntax."
 * Added the old_mode system variable to cause the server to
   revert to certain behaviors present in older versions.
   Currently, this variable affects handling of index hints. See
   Section 13.2.7.2, "Index Hint Syntax."
 * NDB Cluster: Added the --skip-table-check option (short form
   -s) for ndb_restore, which causes the restoration process to
   ignore any changes that may have occurred in table schemas
   after the backup was made. Previously, this was the default
   behavior. (Bug#24363:http://bugs.mysql.com/24363)
   See Section 15.8.3, "ndb_restore --- Restore a Cluster
   Backup," for more information.
 * The server now includes a timestamp in error messages that are
   logged as a result of unhandled signals (such as mysqld got
   signal 11 messages). (Bug#24878:http://bugs.mysql.com/24878)
 * Prefix lengths for columns in SPATIAL indexes no longer can be
   specified. For tables created in older versions of MySQL that
   have SPATIAL indexes containing prefixed columns, dumping and
   reloading the table causes the indexes to be created with no
   prefixes. (The full column width of each column is indexed.)
   (Bug#26794:http://bugs.mysql.com/26794)
 * Added a --no-beep option to mysqladmin. It suppresses the
   warning beep that is emitted by default for errors such as a
   failure to connect to the server.
   (Bug#26964:http://bugs.mysql.com/26964)

Bugs fixed:

 * Incompatible change: INSERT DELAYED statements are not
   supported for MERGE tables, but the MERGE storage engine was
   not rejecting such statements, resulting in table corruption.
   Applications previously using INSERT DELAYED into MERGE table
   will break when upgrading to versions with this fix. To avoid
   the problem, remove DELAYED from such statements.
   (Bug#26464:http://bugs.mysql.com/26464)
 * Creating a table on one SQL node while in single user mode
   caused other SQL nodes to crash.
   (Bug#26997:http://bugs.mysql.com/26997)
 * NDB Cluster: The output from ndb_restore --print_data was
   incorrect for a backup made of a database containing tables
   with TINYINT or SMALLINT columns.
   (Bug#26740:http://bugs.mysql.com/26740)
 * NDB Cluster: After putting the cluster in single user mode
   from one MySQL server, trying to drop an NDB table from a
   second MySQL server also connected to the cluster would cause
   the second MySQL server to hang.
   (Bug#27254:http://bugs.mysql.com/27254)
 * NDB Cluster (Cluster APIs): You can now use the
   ndb_mgm_check_connection() function to determine whether a
   management server is running. See ndb_mgm_check_connection()
   (http://dev.mysql.com/doc/ndbapi/en/ndb-mgm-check-connection.h
   tml).
 * NDB Cluster: mysqld could crash shortly after a data node
   failure following certain DML operations.
   (Bug#27169:http://bugs.mysql.com/27169)
 * NDB Cluster: The management client command node_id STATUS
   displayed the message Node node_id: not connected when node_id
   was not the node ID of a data node.
   (Bug#21715:http://bugs.mysql.com/21715)
   Note:
   The ALL STATUS command in the cluster management client still
   displays status information for data nodes only. This is by
   design. See Section 15.7.2, "Commands in the MySQL Cluster
   Management Client," for more information.
 * NDB Cluster (Cluster APIs): NAND and NOR operations with
   NdbScanFilter did not perform correctly.
   (Bug#24568:http://bugs.mysql.com/24568)
 * NDB Cluster (Cluster Replication): The simultaneous failure of
   a data node and an SQL node could cause replication to fail.
   (Bug#27005:http://bugs.mysql.com/27005)
 * NDB Cluster (Disk Data): Under some circumstances, a data node
   could fail during restart while flushing Disk Data UNDO logs.
   (Bug#27102:http://bugs.mysql.com/27102)
 * NDB Cluster (Disk Data): Creating multiple Disk Data tables
   using different tablespaces could sometimes cause the cluster
   to fail. (Bug#25992:http://bugs.mysql.com/25992)
 * NDB Cluster (Disk Data): ALTER TABLE ... ADD COLUMN ... on a
   Disk Data table moved data for existing non-indexed columns
   from the tablespace into memory.
   (Bug#25880:http://bugs.mysql.com/25880)
 * NDB Cluster (Disk Data): DROP INDEX on a Disk Data table did
   not always move data from memory into the tablespace.
   (Bug#25877:http://bugs.mysql.com/25877)
 * NDB Cluster (Disk Data): When creating a log file group,
   setting INITIAL_SIZE to less than UNDO_BUFFER_SIZE caused data
   nodes to crash. (Bug#25743:http://bugs.mysql.com/25743)
 * NDB Cluster: It was not possible to set LockPagesInMainMemory
   equal to 0. (Bug#27291:http://bugs.mysql.com/27291)
 * NDB Cluster: A race condition could sometimes occur if the
   node acting as master failed while node IDs were still being
   allocated during startup.
   (Bug#27286:http://bugs.mysql.com/27286)
 * NDB Cluster: When a data node was taking over as the master
   node, a race condition could sometimes occur as the node was
   assuming responsibility for handling of global checkpoints.
   (Bug#27283:http://bugs.mysql.com/27283)
 * NDB Cluster: A delete operation using a scan following by an
   insert using a scan could cause a data node to fail.
   (Bug#27203:http://bugs.mysql.com/27203)
 * NDB Cluster: The same failed request from an API node could be
   handled by the cluster multiple times, resulting in reduced
   performance. (Bug#27087:http://bugs.mysql.com/27087)
 * NDB Cluster: The failure of a data node while restarting could
   cause other data nodes to hang or crash.
   (Bug#27003:http://bugs.mysql.com/27003)
 * NDB Cluster: mysqld processes would sometimes crash under high
   load. (Bug#26825:http://bugs.mysql.com/26825)
 * NDB Cluster: When performing an upgrade or downgrade, no
   specific error information was made available when trying to
   upgrade data nodes or SQL nodes before upgrading management
   nodes. (Bug#21296:http://bugs.mysql.com/21296)
 * NDB Cluster: Some values of MaxNoOfTables caused the error Job
   buffer congestion to occur.
   (Bug#19378:http://bugs.mysql.com/19378)
 * NDB Cluster: An invalid pointer was returned following a
   FSCLOSECONF signal when accessing the REDO logs during a node
   restart or system restart.
   (Bug#26515:http://bugs.mysql.com/26515)
 * NDB Cluster (Disk Data): A memory overflow could occur with
   tables having a large amount of data stored on disk, or with
   queries using a very high degree of parallelism on Disk Data
   tables. (Bug#26514:http://bugs.mysql.com/26514)
 * NDB Cluster (Disk Data): Use of a tablespace whose
   INITIAL_SIZE was greater than 1 GB could cause the cluster to
   crash. (Bug#26487:http://bugs.mysql.com/26487)
 * NDB Cluster: Using only the --print_data option (and no other
   options) with ndb_restore caused ndb_restore to fail.
   (Bug#26741:http://bugs.mysql.com/26741)
   This bug was introduced by the fix for
   Bug#14612:http://bugs.mysql.com/14612.
 * NDB Cluster: An infinite loop in an internal logging function
   could cause trace logs to fill up with Unknown Signal type
   error messages and thus grow to unreasonable sizes.
   (Bug#26720:http://bugs.mysql.com/26720)
 * For some operations, system tables in the mysql database must
   be accessed. For example, the HELP statement requires the
   contents of the server-side help tables, and CONVERT_TZ()
   might need to read the time zone tables. However, to perform
   such operations while a LOCK TABLES statement is in effect,
   the server required you to also lock the requisite system
   tables explicitly or a lock error occurred:
      mysql> LOCK TABLE t1 READ;
      Query OK, 0 rows affected (0.02 sec)
      mysql> HELP HELP;
      ERROR 1100 (HY000) at line 4: Table 'help_topic' was not
      locked with LOCK TABLES
   Now, the server implicitly locks the system tables for reading
   as necessary so that you need not lock them explicitly. These
   tables are treated as just described:
      mysql.help_category
      mysql.help_keyword
      mysql.help_relation
      mysql.help_topic
      mysql.proc
      mysql.time_zone
      mysql.time_zone_leap_second
      mysql.time_zone_name
      mysql.time_zone_transition
      mysql.time_zone_transition_type
   If you want to explicitly place a WRITE lock on any of those
   tables with a LOCK TABLES statement, the table must be the
   only one locked; no other table can be locked with the same
   statement. (Bug#9953:http://bugs.mysql.com/9953)
 * For a stored procedure containing a SELECT statement that used
   a complicated join with an ON expression, the expression could
   be ignored during re-execution of the procedure, yielding an
   incorrect result. (Bug#20492:http://bugs.mysql.com/20492)
 * The parser accepted illegal code in SQL exception handlers,
   leading to a crash at runtime when executing the code.
   (Bug#26503:http://bugs.mysql.com/26503)
 * On Windows, debug builds of mysqld could fail with heap
   assertions. (Bug#25765:http://bugs.mysql.com/25765)
 * On Windows, debug builds of mysqlbinlog could fail with a
   memory error. (Bug#23736:http://bugs.mysql.com/23736)
 * SELECT ... INTO OUTFILE with a long FIELDS ENCLOSED BY value
   could crash the server.
   (Bug#27231:http://bugs.mysql.com/27231)
 * The server could hang during binary log rotation.
   (Bug#26079:http://bugs.mysql.com/26079)
 * Increasing the width of a DECIMAL column could cause column
   values to be changed. (Bug#24558:http://bugs.mysql.com/24558)
 * DOUBLE values such as 20070202191048.000000 were being treated
   as illegal arguments by WEEK().
   (Bug#23616:http://bugs.mysql.com/23616)
 * Replication between master and slave would infinitely retry
   binary log transmission where the max_allowed_packet on the
   master was larger than that on the slave if the size of the
   transfer was between these two values.
   (Bug#23775:http://bugs.mysql.com/23775)
 * Setting event_scheduler=1 or event_scheduler=ON caused the
   server to crash if the server had been started with
   --skip-grant-tables. Starting the server with
   --skip-grant-tables now causes event_scheduler to be set to
   DISABLED automatically, overriding any other value that may
   have been set. (Bug#26807:http://bugs.mysql.com/26807)
 * SHOW CREATE EVENT failed to display the STARTS and ENDS
   clauses for an event defined with STARTS NOW(), ENDS NOW(), or
   both. (Bug#26429:http://bugs.mysql.com/26429)
 * Invalid optimization of pushdown conditions for queries where
   an outer join was guaranteed to read only one row from the
   outer table led to results with too few rows.
   (Bug#26963:http://bugs.mysql.com/26963)
 * For INSERT ... ON DUPLICATE KEY UPDATE statements on tables
   containing AUTO_INCREMENT columns, LAST_INSERT_ID() was reset
   to 0 if no rows were successfully inserted or changed. "Not
   changed" includes the case where a row was updated to its
   current values, but in that case, LAST_INSERT_ID() should not
   be reset to 0. Now LAST_INSERT_ID() is reset to 0 only if no
   rows were successfully inserted or touched, whether or not
   touched rows were changed.
   (Bug#27033:http://bugs.mysql.com/27033)
   This bug was introduced by the fix for
   Bug#19978:http://bugs.mysql.com/19978.
 * AFTER UPDATE triggers were not activated by the update part of
   INSERT ... ON DUPLICATE KEY UPDATE statements.
   (Bug#27006:http://bugs.mysql.com/27006)
   This bug was introduced by the fix for
   Bug#19978:http://bugs.mysql.com/19978.
 * For INSERT ... ON DUPLICATE KEY UPDATE statements where some
   AUTO_INCREMENT values were generated automatically for inserts
   and some rows were updated, one auto-generated value was lost
   per updated row, leading to faster exhaustion of the range of
   the AUTO_INCREMENT column.
   (Bug#24432:http://bugs.mysql.com/24432)
   Because the original problem can affect replication (different
   values on master and slave), it is recommended that the master
   and its slaves be upgraded to the current version.
 * For an INSERT statement that should fail due to a column with
   no default value not being assigned a value, the statement
   succeeded with no error if the column was assigned a value in
   an ON DUPLICATE KEY UPDATE clause, even if that clause was not
   used. (Bug#26261:http://bugs.mysql.com/26261)
 * A result set column formed by concatention of string literals
   was incomplete when the column was produced by a subquery in
   the FROM clause. (Bug#26738:http://bugs.mysql.com/26738)
 * When using the result of SEC_TO_TIME() for time value greater
   than 24 hours in an ORDER BY clause, either directly or
   through a column alias, the rows were sorted incorrectly as
   strings. (Bug#26672:http://bugs.mysql.com/26672)
 * If the server was started with --skip-grant-tables, Selecting
   from INFORMATION_SCHEMA tables causes a server crash.
   (Bug#26285:http://bugs.mysql.com/26285)
 * Adding ORDER BY clause to a query on an InnoDB table could
   cause the statement to return no result if the optimizer chose
   a covering index that enabled it to skip the ORDER BY.
   (Bug#24778:http://bugs.mysql.com/24778)
 * IN ((subquery)), IN (((subquery))), and so forth, are
   equivalent to IN (subquery), which is always interpreted as a
   table subquery (so that it is allowed to return more than one
   row). MySQL was treating the "over-parenthesized" subquery as
   a single-row subquery and rejecting it if it returned more
   than one row. This bug primarily affected automatically
   generated code (such as queries generated by Hibernate),
   because humans rarely write the over-parenthesized forms.
   (Bug#21904:http://bugs.mysql.com/21904)
 * For MERGE tables defined on underlying tables that contained a
   short VARCHAR column (shorter than four characters), using
   ALTER TABLE on at least one but not all of the underlying
   tables caused the table definitions to be considered different
   from that of the MERGE table, even if the ALTER TABLE did not
   change the definition.
   (Bug#26881:http://bugs.mysql.com/26881)
 * If a thread previously serviced a connection that was killed,
   excessive memory and CPU use by the thread occurred if it
   later serviced a connection that had to wait for a table lock.
   (Bug#25966:http://bugs.mysql.com/25966)
 * CURDATE() is less than NOW(), either when comparing CURDATE()
   directly (CURDATE() < NOW() is true) or when casting CURDATE()
   to DATE (CAST(CURDATE() AS DATE) < NOW() is true). However,
   storing CURDATE() in a DATE column and comparing col_name <
   NOW() incorrectly yielded false. This is fixed by comparing a
   DATE column as DATETIME for comparisons to a DATETIME
   constant. (Bug#21103:http://bugs.mysql.com/21103)
 * A view on a join is insertable for INSERT statements that
   store values into only one table of the join. However, inserts
   were being rejected if the inserted-into table was used in a
   self-join because MySQL incorrectly was considering the insert
   to modify multiple tables of the view.
   (Bug#25122:http://bugs.mysql.com/25122)
 * Expressions involving SUM(), when used in an ORDER BY clause,
   could lead to out-of-order results.
   (Bug#25376:http://bugs.mysql.com/25376)
 * LOAD DATA INFILE sent an okay to the client before writing the
   binary log and committing the changes to the table had
   finished, thus violating ACID requirements.
   (Bug#26050:http://bugs.mysql.com/26050)
 * Views that used a scalar correlated subquery returned
   incorrect results. (Bug#26560:http://bugs.mysql.com/26560)
 * IF(expr, unsigned_expr, unsigned_expr) was evaluated to a
   signed result, not unsigned. This has been corrected. The fix
   also affects constructs of the form IS [NOT] {TRUE|FALSE},
   which were transformed internally into IF() expressions that
   evaluated to a signed result.
   (Bug#24532:http://bugs.mysql.com/24532)
   For existing views that were defined using IS [NOT]
   {TRUE|FALSE} constructs, there is a related implication. The
   definitions of such views were stored using the IF()
   expression, not the original construct. This is manifest in
   that SHOW CREATE VIEW shows the transformed IF() expression,
   not the original one. Existing views will evaluate correctly
   after the fix, but if you want SHOW CREATE VIEW to display the
   original construct, you must drop the view and re-create it
   using its original definition. New views will retain the
   construct in their definition.
 * BENCHMARK() did not work correctly for expressions that
   produced a DECIMAL result.
   (Bug#26093:http://bugs.mysql.com/26093)
 * For MEMORY tables, extending the length of a VARCHAR column
   with ALTER TABLE might result in an unusable table.
   (Bug#26080:http://bugs.mysql.com/26080)
 * For some values of the position argument, the INSERT()
   function could insert a NUL byte into the result.
   (Bug#26281:http://bugs.mysql.com/26281)
 * Creating a table with latin characters in the name caused the
   output of SHOW FULL TABLES to have ERROR for the table type.
   (Bug#25081:http://bugs.mysql.com/25081)
 * Inserting utf8 data into a TEXT column that used a single-byte
   character set could result in spurious warnings about
   truncated data. (Bug#25815:http://bugs.mysql.com/25815)
 * EXPLAIN EXTENDED did not show WHERE conditions that were
   optimized away. (Bug#22331:http://bugs.mysql.com/22331)
 * INSERT DELAYED statements inserted incorrect values into BIT
   columns. (Bug#26238:http://bugs.mysql.com/26238)
 * For expr IN(value_list), the result could be incorrect if
   BIGINT UNSIGNED values were used for expr or in the value
   list. (Bug#19342:http://bugs.mysql.com/19342)
 * When a TIME_FORMAT() expression was used as a column in a
   GROUP BY clause, the expression result was truncated.
   (Bug#20293:http://bugs.mysql.com/20293)
 * For SUBSTRING() evaluation using a temporary table, when
   SUBSTRING() was used on a LONGTEXT column, the max_length
   metadata value of the result was incorrectly calculated and
   set to 0. Consequently, an empty string was returned instead
   of the correct result.
   (Bug#15757:http://bugs.mysql.com/15757)
 * Use of a GROUP BY clause that referred to a stored function
   result together with WITH ROLLUP caused incorrect results.
   (Bug#25373:http://bugs.mysql.com/25373)
 * Use of a subquery containing GROUP BY and WITH ROLLUP caused a
   server crash. (Bug#26830:http://bugs.mysql.com/26830)
 * Use of a subquery containing a UNION with an invalid ORDER BY
   clause caused a server crash.
   (Bug#26661:http://bugs.mysql.com/26661)
 * In certain cases it could happen that deleting a row corrupted
   an RTREE index. This affected indexes on spatial columns.
   (Bug#25673:http://bugs.mysql.com/25673)
 * SSL connections failed on Windows.
   (Bug#26678:http://bugs.mysql.com/26678)
 * Added support for --debugger=dbx for mysql-test-run.pl and
   fixed support for --debugger=devenv, --debugger=DevEnv, and
   --debugger=/path/to/devenv.
   (Bug#26792:http://bugs.mysql.com/26792)
 * X() IS NULL and Y() IS NULL comparisons failed when X() and
   Y() returned NULL. (Bug#26038:http://bugs.mysql.com/26038)
 * UNHEX() IS NULL comparisons failed when UNHEX() returned NULL.
   (Bug#26537:http://bugs.mysql.com/26537)
 * The REPEAT() function did not allow a column name as the count
   parameter. (Bug#25197:http://bugs.mysql.com/25197)
 * On 64-bit Windows, large timestamp values could be handled
   incorrectly. (Bug#26536:http://bugs.mysql.com/26536)
 * In some error messages, inconsistent format specifiers were
   used for the translations in different languages. comp_err
   (the error message compiler) now checks for mismatches.
   (Bug#26571:http://bugs.mysql.com/26571)
 * On Windows, the server exhibited a file-handle leak after
   reaching the limit on the number of open file descriptors.
   (Bug#25222:http://bugs.mysql.com/25222)
 * A reference to a non-existent column in the ORDER BY clause of
   an UPDATE ... ORDER BY statement could cause a server crash.
   (Bug#25126:http://bugs.mysql.com/25126)
 * mysqlimport used a variable of the wrong type for the
   --use-threads option, which could cause a crash on some
   architectures. (Bug#23814:http://bugs.mysql.com/23814)
 * A multiple-row delayed insert with an auto increment column
   could cause duplicate entries to be created on the slave in a
   replication environment.
   (Bug#25507:http://bugs.mysql.com/25507,
   Bug#26116:http://bugs.mysql.com/26116)
 * Using mysqlbinlog on a binary log would crash if there were a
   large number of row-based events related to a single
   statement. (Bug#25628:http://bugs.mysql.com/25628)
 * Duplicating the usage of a user variable in a stored procedure
   or trigger would not be replicated correctly to the slave.
   (Bug#25167:http://bugs.mysql.com/25167)
 * User defined variables used within stored procedures and
   triggers are not replicated correctly when operating in
   statement-based replication mode.
   (Bug#20141:http://bugs.mysql.com/20141,
   Bug#14914:http://bugs.mysql.com/14914)
 * Loading data using LOAD DATA INFILE may not replicate
   correctly (due to character set incompatibilities) if the
   character_set_database variable is set before the data is
   loaded. (Bug#15126:http://bugs.mysql.com/15126)
 * DROP TRIGGER statements would not be filtered on the slave
   when using the replication-wild-do-table option.
   (Bug#24478:http://bugs.mysql.com/24478)
 * SHOW ENGINE MUTEX STATUS failed to produce an Unknown table
   engine error. (Bug#24392:http://bugs.mysql.com/24392)
   See Section 13.5.4.12, "SHOW ENGINE Syntax."
 * MySQL would not compile when configured using
   --without-query-cache.
   (Bug#25075:http://bugs.mysql.com/25075)
 * When using certain server SQL modes, the mysql.proc table was
   not created by mysql_install_db. In addition, the creation of
   this and other MySQL system tables was not checked for by
   mysql-test-run.pl. (Bug#23669:http://bugs.mysql.com/23669,
   Bug#20166:http://bugs.mysql.com/20166)
 * It was not possible to use XPath keywords as tag names for
   expressions used in the ExtractValue() function.
   (Bug#24747:http://bugs.mysql.com/24747)
 * VIEW restrictions were applied to SELECT statements after a
   CREATE VIEW statement failed, as though the CREATE had
   succeeded. (Bug#25897:http://bugs.mysql.com/25897)
 * An INSERT trigger invoking a stored routine that inserted into
   a table other than the one on which the trigger was defined
   would fail with a Table '...' doesn't exist referring to the
   second table when attempting to delete records from the first
   table. (Bug#21825:http://bugs.mysql.com/21825)
 * A stored procedure that made use of cursors failed when the
   procedure was invoked from a stored function.
   (Bug#25345:http://bugs.mysql.com/25345)
 * When nesting stored procedures within a trigger on a table, a
   false dependency error was thrown when one of the nested
   procedures contained a DROP TABLE statement.
   (Bug#22580:http://bugs.mysql.com/22580)
 * When attempting to call a stored procedure creating a table
   from a trigger on a table tbl in a database db, the trigger
   failed with ERROR 1146 (42S02): Table 'db.tbl' doesn't exist.
   However, the actual reason that such a trigger fails is due to
   the fact that CREATE TABLE causes an implicit COMMIT, and so a
   trigger cannot invoke a stored routine containing this
   statement. A trigger which does so now fails with ERROR 1422
   (HY000): Explicit or implicit commit is not allowed in stored
   function or trigger, which makes clear the reason for the
   trigger's failure. (Bug#18914:http://bugs.mysql.com/18914)
 * Local variables in stored routines or triggers, when declared
   as the BIT type, were interpreted as strings.
   (Bug#12976:http://bugs.mysql.com/12976)
 * When a stored routine attempted to execute a statement
   accessing a nonexistent table, the error was not caught by the
   routine's exception handler.
   (Bug#8407:http://bugs.mysql.com/8407,
   Bug#20713:http://bugs.mysql.com/20713)
 * NOW() returned the wrong value in statements executed at
   server startup with the --init-file option.
   (Bug#23240:http://bugs.mysql.com/23240)
 * Instance Manager did not remove the angel PID file on a clean
   shutdown. (Bug#22511:http://bugs.mysql.com/22511)
 * Setting the slow_query_log_file system variable caused log
   output to go tothe general log, not the slow query log.
   (Bug#23225:http://bugs.mysql.com/23225)
 * The server could crash if two or more threads initiated query
   cache resize operation at moments very close in time.
   (Bug#23527:http://bugs.mysql.com/23527)
 * The conditions checked by the optimizer to allow use of
   indexes in IN predicate calculations were unnecessarily tight
   and were relaxed. (Bug#20420:http://bugs.mysql.com/20420)
 * Several deficiencies in resolution of column names for INSERT
   ... SELECT statements were corrected.
   (Bug#25831:http://bugs.mysql.com/25831)
 * Indexes on TEXT columns were ignored when ref accesses were
   evaluated. (Bug#25971:http://bugs.mysql.com/25971)
 * The update columns for INSERT ... SELECT ... ON DUPLICATE KEY
   UPDATE could be assigned incorrect values if a temporary table
   was used to evaluate the SELECT.
   (Bug#16630:http://bugs.mysql.com/16630)
 * A user-defined variable could be assigned an incorrect value
   if a temporary table was employed in obtaining the result of
   the query used to determine its value.
   (Bug#24010:http://bugs.mysql.com/24010)
 * Queries that used a temporary table for the outer query when
   evaluating a correlated subquery could return incorrect
   results. (Bug#23800:http://bugs.mysql.com/23800)
 * For index reads, the BLACKHOLE engine did not return
   end-of-file (which it must because BLACKHOLE tables contain no
   rows), causing some queries to crash.
   (Bug#19717:http://bugs.mysql.com/19717)
 * A query of type index_merge, and with a WHERE clause having
   the form WHERE indexed_column_1=value_1 OR
   indexed_column_2=value_2 on a partitioned table caused the
   server to crash. (Bug#26117:http://bugs.mysql.com/26117)