Incompatible Change: A change has been made to the way that the server handles prepared statements. This affects prepared statements processed at the SQL level (using the PREPARE statement) and those processed using the binary client-server protocol (using the mysql_stmt_prepare() C API function).

Previously, changes to metadata of tables or views referred to in a prepared statement could cause a server crash when the statement was next executed, or perhaps an error at execute time with a crash occurring later. For example, this could happen after dropping a table and recreating it with a different definition.

Now metadata changes to tables or views referred to by prepared statements are detected and cause automatic repreparation of the statement when it is next executed. Metadata changes occur for DDL statements such as those that create, drop, alter, rename, or truncate tables, or that analyze, optimize, or repair tables. Repreparation also occurs after referenced tables or views are flushed from the table definition cache, either implicitly to make room for new entries in the cache, or explicitly due to FLUSH TABLES.

Repreparation is automatic, but to the extent that it occurs, performance of prepared statements is diminished.

Table content changes (for example, with INSERT or UPDATE) do not cause repreparation, nor do SELECT statements.

An incompatibility with previous versions of MySQL is that a prepared statement may now return a different set of columns or different column types from one execution to the next. For example, if the prepared statement is SELECT * FROM t1, altering t1 to contain a different number of columns causes the next execution to return a number of columns different from the previous execution.

Older versions of the client library cannot handle this change in behavior. For applications that use prepared statements with the new server, an upgrade to the new client library is strongly recommended.

Along with this change to statement repreparation, the default value of the table_definition_cache system variable has been increased from 128 to 256. The purpose of this increase is to lessen the chance that prepared statements will need repreparation due to referred-to tables/views having been flushed from the cache to make room for new entries.

A status variable, Com_stmt_reprepare, has been introduced to track the number of repreparations. (Bug#27420, Bug#27430, Bug#27690)

Important Change: Some changes were made to CHECK TABLE ... FOR UPGRADE and REPAIR TABLE with respect to detection and handling of tables with incompatible .frm files (files created with a different version of the MySQL server). These changes also affect mysqlcheck because that program uses CHECK TABLE and REPAIR TABLE, and thus also mysql_upgrade because that program invokes mysqlcheck.

(Bug#36055)

mysql_upgrade now has a --tmpdir option to enable the location of temporary files to be specified. (Bug#36469)

mysqldump now adds the LOCAL qualifier to the FLUSH TABLES statement that is sent to the server when the --master-data option is enabled. This prevents the FLUSH TABLES statement from replicating to slaves, which is disadvantageous because it would cause slaves to block while the statement executes. (Bug#35157)

See also Bug#38303.

Important Change: The server no longer issues warnings for truncation of excess spaces for values inserted into CHAR columns. This reverts a change in the previous release that caused warnings to be issued. (Bug#30059)

Replication: CREATE PROCEDURE and CREATE FUNCTION statements containing extended comments were not written to the binary log correctly, causing parse errors on the slave. (Bug#36570)

See also Bug#32575.

Replication: When flushing tables, there was a slight chance that the flush occurred between the processing of one table map event and the next. Since the tables were opened one by one, subsequent locking of tables would cause the slave to crash. This problem was observed when replicating NDBCLUSTER or InnoDB tables, when executing multi-table updates, and when a trigger or a stored routine performed an (additional) insert on a table so that two tables were effectively being inserted into in the same statement. (Bug#36197)

Replication: CREATE VIEW statements containing extended comments were not written to the binary log correctly, causing parse errors on the slave. Now, all comments are stripped from such statements before being written to the binary log. (Bug#32575)

See also Bug#36570.

On Windows 64-bit systems, temporary variables of long types were used to store ulong values, causing key cache initialization to receive distorted parameters. The effect was that setting key_buffer_size to values of 2GB or more caused memory exhaustion to due allocation of too much memory. (Bug#36705)

Multiple-table UPDATE statements that used a temporary table could fail to update all qualifying rows or fail with a spurious duplicate-key error. (Bug#36676)

A REGEXP match could return incorrect rows when the previous row matched the expression and used CONCAT() with an empty string. (Bug#36488)

mysqltest ignored the value of --tmpdir in one place. (Bug#36465)

When updating an existing instance (for example, from MySQL 5.0 to 5.1, or 5.1 to 6.0), the Instance Configuration Wizard unnecessarily prompted for a root password when there was an existing root password. (Bug#36305)

Conversion of a FLOAT ZEROFILL value to string could cause a server crash if the value was NULL. (Bug#36139)

On Windows, the installer attempted to use JScript to determine whether the target data directory already existed. On Windows Vista x64, this resulted in an error because the installer was attempting to run the JScript in a 32-bit engine, which wasn't registered on Vista. The installer no longer uses JScript but instead relies on a native WiX command. (Bug#36103)

mysqltest was performing escape processing for the --replace_result command, which it should not have been. (Bug#36041)

An error in calculation of the precision of zero-length items (such as NULL) caused a server crash for queries that employed temporary tables. (Bug#36023)

For EXPLAIN EXTENDED, execution of an uncorrelated IN subquery caused a crash if the subquery required a temporary table for its execution. (Bug#36011)

The MERGE storage engine did a table scan for SELECT COUNT(*) statements when it could calculate the number of records from the underlying tables. (Bug#36006)

The server crashed inside NOT IN subqueries with an impossible WHERE or HAVING clause, such as NOT IN (SELECT ... FROM t1, t2, ... WHERE 0). (Bug#36005)

The Event Scheduler was not designed to work under the embedded server. It is now disabled for the embedded server, and the event_scheduler system variable is not displayed. (Bug#35997)

Grouping or ordering of long values in unindexed BLOB or TEXT columns with the gbk or big5 character set crashed the server. (Bug#35993)

SET GLOBAL debug='' resulted in a Valgrind warning in DbugParse(), which was reading beyond the end of the control string. (Bug#35986)

The “prefer full scan on clustered primary key over full scan of any secondary key” optimizer rule introduced by Bug#26447 caused a performance regression for some queries, so it has been disabled. (Bug#35850)

The server ignored any covering index used for ref access of a table in a query with ORDER BY if this index was incompatible with the ORDER BY list and there was another covering index compatible with this list. As a result, suboptimal execution plans were chosen for some queries that used an ORDER BY clause. (Bug#35844)

mysql_upgrade did not properly update the mysql.event table. (Bug#35824)

An incorrect error and message was produced for attempts to create a MyISAM table with an index (.MYI) file name that was already in use by some other MyISAM table that was open at the same time. For example, this might happen if you use the same value of the INDEX DIRECTORY table option for tables belonging to different databases. (Bug#35733)

Enabling the read_only system variable while autocommit mode was enabled caused SELECT statements for transactional storage engines to fail. (Bug#35732)

The combination of GROUP_CONCAT(), DISTINCT, and LEFT JOIN could crash the server when the right table is empty. (Bug#35298)

Some binaries produced stack corruption messages due to being built with versions of bison older than 2.1. Builds are now created using bison 2.3. (Bug#34926)

The log_output system variable could be set to an illegal value. (Bug#34820)

On Windows 64-bit builds, an apparent compiler bug caused memory overruns for code in innobase/mem/*. Removed optimizations so as not to trigger this problem. (Bug#34297)

Several additional configuration scripts in the BUILD directory now are included in source distributions. These may be useful for users who wish to build MySQL from source. (See Section 2.3.3, “Installing from the Development Source Tree”, for information about what they do.) (Bug#34291)

Executing a FLUSH PRIVILEGES statement after creating a temporary table in the mysql database with the same name as one of the MySQL system tables caused the server to crash.

(Bug#33275)

UNION constructs cannot contain SELECT ... INTO except in the final SELECT. However, if a UNION was used in a subquery and an INTO clause appeared in the top-level query, the parser interpreted it as having appeared in the UNION and raised an error. (Bug#32858)

Assignment of relative path names to general_log_file or slow_query_log_file did not always work. (Bug#32748)

The mysql.servers table was not created during installation on Windows. (Bug#28680, Bug#32797)

The jp test suite was not working. (Bug#28563)

The internal init_time() library function was renamed to my_init_time() to avoid conflicts with external libraries. (Bug#26294)

The parser used signed rather than unsigned values in some cases that caused legal lengths in column declarations to be rejected. (Bug#15776)