[This is part 3 of the announcement] * JSON: When a JSON value consisted of a large sub-document wrapped in many levels of JSON arrays, objects, or both, serialization of the JSON value sometimes required an excessive amount time to complete. (Bug #23031146)
* JSON: When a NULL value existed in a JSON column, the result from a query using GROUP_CONCAT() together with the ORDER BY clause was not always correct. (Bug #22992666) * JSON: The internal rapid_json_handler used its own data structures to represent a partially-built DOM; these had to be converted into a Json_dom graph before returning the result. Now this handler builds the graph directly, which reduces the amount of work required to build it, and thus to parse a JSON document. (Bug #22900110) * JSON: The internal function Item_func_case::val_json() did not always set the null value flag as expected when a CASE expression evaluated to NULL, leading to an assertion in debug builds of the server. (Bug #22887227) * JSON: The SUM() function truncated decimal values extracted from JSON documents, producing an integer result. (Bug #84935, Bug #25530204) * JSON: A JSON document that contained a double value slightly greater than the maximum value that can be represented by a double silently replaced it with zero instead of rejecting the value and raising an error. Such values are now handled correctly in MySQL JSON documents. The underlying issue was traced to a problem with RapidJSON, which has been reported to that library's developers as Issue #849 (https://github.com/miloyip/rapidjson/issues/849). (Bug #84891, Bug #25518504) * JSON: The JSON_SEARCH() and JSON_CONTAINS_PATH() functions did not work when the one_or_all argument was specified using UTF-16 encoding. For both of these functions, this argument is now converted to utfmb4 if need be before its value is checked. (Bug #84880, Bug #22516960) * JSON: The JSON_UNQUOTE() function did not work with strings that used UTF-16 encoding. Now these strings are converted to utfmb4 internally before being processed. (Bug #84878, Bug #25516881) * JSON: Updating the same JSON column in a single statement could cause incorrect values to be written into the table. This occurred when the second update overwrote the column value with a subset of itself. An example of such a statement is shown here: UPDATE t SET col = JSON_ARRAY(value), col = col->'$[0]'; (Bug #84694, Bug #25461627) * JSON: The functions JSON_QUOTE() and JSON_UNQUOTE() did not work correctly with multibyte character sets such as utf8mb4. (Bug #84680, Bug #25455065) References: See also: Bug #77234, Bug #21193273. * JSON: The error message for Error 3152 ER_JSON_USED_AS_KEY has been changed from JSON column '%s' cannot be used in key specification to the less confusing and more accurate JSON column '%s' supports indexing only via generated columns on a specified JSON path. (Bug #81364, Bug #23274244) * JSON: Internal tests for MySQL JSON functionality ran out of stack space on some platforms when run against a debug-enabled server. Because timely checks were not made for stack usage, the server did not detect this situation, leading to a server exit. The fix for this issue is twofold: + Stack overrun checks are now made before attempting to serialize a nested array or object, so that the operation fails gracefully when processing deeply nested JSON documents, rather than causing an exit. + Serialization of JSON documents has been reorganized so that it requires less use of the stack when compiled without optimization. (Bug #81083, Bug #23106330) * MySQL did not compile with GCC 7. (Bug #25643811) * The (undocumented) WINDOWS_RUNTIME_MD CMake option has been removed. (Bug #25611359) * If --skip-innodb or one of its variants was used, a spurious warning about avoid_temporal_upgrade was generated. (Bug #25573578) * mysqld_safe failed to restart the server if a PID_FILE.shutdown file was present. (Bug #25572504) References: This issue is a regression of: Bug #11751149. * For Debian/Ubuntu packages, user-defined collation files could be overwritten during MySQL upgrades. Charset files are now marked as conffiles so that user customizations generate a prompt during upgrades whether to overwrite them. (Bug #25525628, Bug #84761) * For CREATE TABLE statements that specified the table name with a database qualifier and included a DATA DIRECTORY or INDEX DIRECTORY option, an error occurred if there was no default database. (Bug #25514146, Bug #84861) * referenced_table_schema and referenced_table_name field values in the mysql.foreign_keys data dictionary table were not stored in lowercase when lower_case_table_names was enabled. (Bug #25495714) * Starting the server with performance_schema_digests_size=1 caused an abnormal exit. (Bug #25492129, Bug #84786) * For clients linked against yaSSL, connections became invalid when a read timeout occurred, rather than retrying the read. (Bug #25444075) * MySQL compilation in different directories produced different builds to leakage of absolute paths into debug information and __FILE__. (Bug #25436469) * Calculations for UCA 9.0.0 collations were inefficient for tailoring rules containing contraction characters. (Bug #25426632) * A negative internal connection timeout value caused the connection to go idle and abort. A negative value now causes the connection to block indefinitely in the absence of I/O. (Bug #25408557) * A mutex Performance Schema name was too long and produced a warning at server startup. (Bug #25406915) * MySQL failed to compile on some platforms with -DWITH_LIBWRAP=ON. CMake support now checks whether tcpd.h has proper function prototypes. (Bug #25395543, Bug #84495) * An index defined on a data dictionary table column exceeded the maximum index key length when the instance was initiated with innodb_page_size=4k. (Bug #25384527) * mysqld_safe did not check whether the directory named by the --basedir option existed. (Bug #25365194) * Configuring CMake with -G ninja resulted in build output that was inappropriate for build platforms other than Xcode or Visual Studio. (Bug #25358460) * mysqld_safe failed if the error log file named by the --log-error option was a FIFO. (Bug #25356221, Bug #84427) * For prepared statements, an alias within a subquery or derived table might cause incorrect behavior during statement execution if another alias depended on it. (Bug #25343335, Bug #84398, Bug #25171608) * mysqld_safe could fail if the --datadir option value ended with a / character. (Bug #25319457) * A recent change to mysqld_safe caused the mysql.server script to be unable to start it if the base directory was specified as an absolute path that differed from the compiled-in default absolute path. (Bug #25319392, Bug #84263) * The connection_control plugin failed to compile if the Performance Schema was disabled. (Bug #25308357, Bug #84304) * Passwords did not expire correctly for accounts created using MySQL Workbench. (Bug #25299309) * For System V init scripts for RPMs, the [mysqld] option-file section was being ignored for some options, such as pid-file. (Bug #25287707, Bug #84172) * Init scripts failed to launch mysqld_safe if a non-default base directory was used. (Bug #25261472, Bug #84219) * CMake now detects whether a GCC 5.3.0 loop optimization bug occurs and attempts a workaround if so. (Bug #25253540) * mysqld_safe --no-defaults did not work (inadvertent consequence of an earlier bug fix). (Bug #25244898, Bug #84173) * Semicolon (;) characters within or between statements could cause distinct digests to be generated from identical statements. (Bug #25244533, Bug #83253) * Components could not register services without referring to their private implementation. A SERVICE_IMPLEMENTATION(component,service) macro now enables this to be avoided. (Bug #25238906) * Certain stored functions, if used in a query WHERE clause, could be handled using Index Condition Pushdown (which should not happen), resulting in a server exit. (Bug #25196653, Bug #25174454) * For ai_ci collations based on Unicode Collation Algorithm 9.0.0, accented characters that compare equal were treated as different by LIKE comparisons. (Bug #25167284, Bug #83999) * For a client linked against libmysqlclient, invalid memory access could occur during use of prepared statements. (Bug #25164932) * Some Linux startup scripts did not process the datadir setting correctly. (Bug #25159791) * If a character set is specified for a column of a partitioned table, a segmentation fault could occur while upgrading. (Bug #25153261) * LOAD DATA failed to accept multibyte characters that followed an escape sequence. (Bug #25147988, Bug #83950) * The fix for Bug #25088048 caused the command used by mysqld_safe to start the MySQL server to no longer include the mysqld path. (Bug #25144379) References: This issue is a regression of: Bug #25088048. * For UCA collations, LIKE comparisons against a pattern that ended with the escape character returned incorrect results. (Bug #25140629, Bug #83930) * The default character set and collation were used instead of the character set and collation defined in db.opt file when upgrading the schema from MySQL 5.7 to MySQL 8.0. (Bug #25139901) * Instead of updating the data dictionary cache at the end of DDL statements using separate function calls, the data dictionary cache is now updated as part of transaction commit. (Bug #25095798, Bug #83818) * CREATE TABLE now requires the FILE privilege if DATA DIRECTORY or INDEX DIRECTORY is specified explicitly as a table or partition option. ALTER TABLE requires the FILE privilege if either option is specified explicitly as a partition option (it ignores them if specified as table options). (Bug #25092566) * There were some differences between ICU order and MySQL Unicode collations for Hungarian contractions and ligatures. MySQL now follows ICU order. (Bug #25090543) * The data dictionary failed to initialize and start when using a binary collation. The query string generated to create the data dictionary schema did not add quotes to the binary collation name. (Bug #25054104, Bug #83706) * Executing a stored procedure containing a query that accessed a view could allocate memory that was not freed until the session ended. (Bug #25053286) * Compilation on FreeBSD 11 failed attempting to check MAP_NORESERVE, which is no longer defined. (Bug #25048128, Bug #83689) * mysql-test-run.pl now checks whether the TSAN_OPTIONS environment variable is set. If so, the value is taken as the path name of a file containing ThreadSanitizer suppressions (errors to be ignored during test runs). Additionally, the --sanitize option now causes mysql-test-run.pl to scan the server error logs for ThreadSanizer messages. (Bug #24970905, Bug #83601) * After starting the server with a non-zero --lower-case-table-names setting, an assertion was raised when USE INFORMATION_SCHEMA; was the first instruction from the first client. (Bug #24963580) * For case-insensitive Unicode collations, the various space characters did not hash to the same value, resulting in incorrect comparisons between them. (Bug #24956750) * Dictionary clients now track uncommitted dictionary objects that are being modified by DDL statements, making uncommitted changes by the same session visible to a dictionary client without affecting the dictionary object cache prior to commit time. (Bug #24956365, Bug #83548) * Simultaneous dictionary object cache misses were not handled correctly. An object retrieved by one thread could be evicted from the cache before a waiting thread attempted to access it. (Bug #24949179) * CMake now avoids configuring the -fexpensive-optimizations option for GCC versions for which the option triggers faulty shift-or optimizations. (Bug #24947597, Bug #83517) * CMake support was added for compiling using the -std=c++03 option under Developer Studio 12.5. This is now used rather than stlport by default. (Bug #24947136, Bug #83512, Bug #25229424) * NCHAR and NATIONAL CHAR are synonyms, but CAST(expr AS NCHAR) succeeded, whereas CAST(expr AS NATIONAL CHAR) did not. Now both work. (Bug #24934161) * OEL RPM packages now better detect which platforms have multilib support (for which 32-bit and 64-bit libraries can be installed). Thanks to Alexey Kopytov for the patch. (Bug #24925181, Bug #83457) * OEL RPM packages now better detect which platforms do not have multilib support (for which 32-bit and 64-bit libraries can be installed). Thanks to Alexey Kopytov for the patch. (Bug #24916428, Bug #83428) * The LOCATE() function returned NULL if the substr or str argument was NULL, but not if the pos argument was NULL. Now it returns NULL if any argument is NULL. (Bug #24911350, Bug #83427) * Bit operations could cause a server exit to occur if argument nullability was mishandled. (Bug #24910958, Bug #24930038, Bug #24930829) * The Block Nested Loop algorithm could allocate too much memory during query execution. (Bug #24909223) * Information about building MySQL 5.6 compatibility libraries in the MySQL 5.7 and higher .spec file is needed only for building libmysqlclient and libmysqld. Information about building the InnoDB memcached plugin was removed. (Bug #24908345, Bug #83409) * Incorrect updating of view metadata could raise an assertion. (Bug #24834622) * Initialization of the keyring_okv plugin failed if the STANDBY_SERVER setting was missing from the okvclient.ora configuration file, effectively making this a mandatory setting. STANDBY_SERVER is now optional. (Bug #24816271) * Data dictionary objects acquired by Dictionary_client::acquire_uncached() are now owned by the current auto-releaser instead of the caller. Also, acquisition of TABLE_SHARE view objects is performed by acquire() and clone() instead of acquire_uncached(). (Bug #24813358, Bug #83296) * Privilege checking could be incorrect for a derived table used within a multiple-table UPDATE invoked within a stored procedure or view object, for the second or subsequent execution of the object, if the derived table was merged into the outer query. (Bug #24810564) * The Created_tmp_tables status variable was incremented in some cases when no temporary table was created. (Bug #24808970, Bug #83287) * An in-place upgrade from MySQL 5.7 to MySQL 8.0 failed if parsing of a stored routine body failed while migrating the routine. Now a warning is reported and the routine is created without parsing its body. Also, warnings are now reported when dependency resolution fails for a view during view migration. (Bug #24805140, Bug #83275) * For compilation, mysql_upgrade is dependent on the dynamically generated sql_commands_system_tables_data_fix.h file, but a missing dependency could cause that file not to be generated. (Bug #24802377, Bug #83272) * A race condition between transactions accessing the access-control list (ACL) cache and use of the ACL cache to populate INFORMATION_SCHEMA tables could cause a server exit. (Bug #24786029) * For debug builds, EXPLAIN or DESCRIBE for a table with a database or table name longer than the maximum permitted length raised an assertion rather than displaying an appropriate error. (Bug #24751177, Bug #83114) * The implementation of several INFORMATION_SCHEMA tables as views on data dictionary tables introduced a number of native SQL functions intended only for internal use by the server, but they could be invoked by users. Those functions now produce an error if invoked by users. (Bug #24749248, Bug #83189) * CASE, COALESCE(), IF() and IFNULL() could merge a mix of signed and unsigned arguments incorrectly and produce an incorrect result type. (Bug #24733658, Bug #83148) * Connections from a client to a server with SSL enabled succeeded even if --ssl-mode had a value of VERIFY_CA or VERIFY_IDENTITY and the client did not provide a CA certificate. (Bug #24732452, Bug #23189252) * Manual creation of a directory in the data directory resulted in USE dir_name succeeding even though the directory was not registered as a database in the data dictionary. (Bug #24732194, Bug #83140) * If InnoDB statistics were incorrect, FOUND_ROWS() could return 1 even when the previous SELECT returned no rows. (Bug #24714857, Bug #83110) * ALTER TABLE on a temporary table could raise an assertion if a nontemporary table with the same name existed. (Bug #24713918, Bug #83117) * CMake now sets -DWITH_NUMA=ON for Debian platforms where possible. (Bug #24689101) * To better provide atomic file creation, Debian packaging scripts now use the coreutils install command rather than touch, chmod, and chown. (Bug #24688682) * Enabling the DISABLE_SHARED CMake option caused compilation failure. (Bug #24687701, Bug #83039) References: This issue is a regression of: Bug #24481181. * The BIN(), OCT(), and HEX() functions could mishandle values of the BIT data type. (Bug #24686658, Bug #83031) * For debug builds, an invalid utf8 character in the comment of a CREATE EVENT or ALTER EVENT statement raised an assertion. This now produces an ER_INVALID_CHARACTER_STRING error. (Bug #24679962) * The encoding for the utf8 character set permitted characters between U+D800 and U+DFFF as valid, though they are reserved for surrogate pairs and do not directly represent characters. They are now considered invalid. (Bug #24672415) * A query could produce incorrect results if the WHERE clause contained a dependent subquery, the table had a secondary index on the columns in the select list followed by the columns in the subquery, and GROUP BY or DISTINCT permitted the query to use a Loose Index Scan. (Bug #24671968, Bug #83005) * After a RENAME TABLE operation on a table with a foreign key, the generated constraint name was not updated in TABLE_CONSTRAINTS. (Bug #24666169) * On macOS, CMake AddressSanitizer support did not work. (Bug #24661626, Bug #82976) References: This issue is a regression of: Bug #23759968. * Compilation failed on macOS 10.11 with Xcode 8.0. (Bug #24661523, Bug #82975) * Index hints applied to invisible indexes produced no error. (Bug #24660093, Bug #82960) * In some cases, INFORMATION_SCHEMA.KEY_COLUMN_USAGE did not report data for all foreign key constraints. (Bug #24655803, Bug #82961) * REPLACE() on large strings could be slow and unkillable. (Bug #24652792) * The DebugPrintTest and DebugPrintDeathTest unit tests did not handle divide-by-zero testing properly on the Aarch64 platform. Thanks to Alexey Kopytov for the patch. (Bug #24624555, Bug #82889) * Compilation on FreeBSD using GCC 6 did not work. (Bug #24619561, Bug #82922) * Changes made to mysqld_safe in recent MySQL releases require the --ledir, --mysqld, --mysqld-version options to be specified on the command line; they can no longer be specified in option files. This could cause failure of init scripts that invoke mysqld_safe. Such scripts now pass the value of the MYSQLD_OPTS environment variable as the first command-line argument to mysqld_safe, with the value set to such command line-only mysqld_safe option values as may be required. On platforms that use /etc/sysconfig/mysqld, the MYSQLD_OPTS value can be set in that file with a line such as this: MYSQLD_OPTS=" --ledir=/mysqld_ledir --mysqld=my_wrapper " The value of MYSQLD_OPTS can also include mysqld options for mysqld_safe to pass to mysqld. (Bug #24619033, Bug #82920) References: This issue is a regression of: Bug #24464380, Bug #24483092, Bug #25088048, Bug #25378439, Bug #25378565. * SET PERSIST innodb_buffer_pool_size = value stored the original value of innodb_buffer_pool_size to mysqld-auto.cnf, not the new value. (Bug #24613005, Bug #82905) * For SLES packages, a typo in the installation script postamble prevented some cleanup from occurring. (Bug #24605300, Bug #82389) * Some messages written by the server to the error log while upgrading the data directory were missing the standard timestamp and process ID information. (Bug #24600054, Bug #82874) * It was possible for the session values of the Last_query_cost and Last_query_partial_plans status variables to be accessed before they had been initialized. (Bug #24596263) * Warnings occurring during CREATE TABLE ... SELECT could cause a server exit. (Bug #24595992) * For LOAD DATA used to insert data into an updateable view, the check to verify whether a column is actually updatable was missing. (Bug #24595937) * The server could dereference a null pointer when a deterministic function returning LONGTEXT was used in a subquery. (Bug #24595581) * A view altered with ALTER VIEW might be dropped if the statement failed with an error. (Bug #24594140) * Conversion of JSON documents to string could be slow if the document was large and contained many signed integers. (Bug #24586888) * The data dictionary can contain entries for temporary tables (names beginning with #sql). These tables were exposed to INFORMATION_SCHEMA queries and through SHOW statements. This could cause mysqldump and mysqlpump to fail when they attempted to dump such tables. Temporary tables are now hidden to INFORMATION_SCHEMA queries and SHOW statements. (Bug #24580599, Bug #24571427) * When a view was defined, the character set of the definition was not considered. If this differed from the default character set, table names used in the view might not be recognized, causing the view to be marked invalid. For debug builds, an assertion could be raised if an error occurred while parsing an ALTER VIEW statement and the diagnostics area was empty. (Bug #24580586) * For LOAD DATA statements, input data with too many column values produced only a warning, rather than an error as in MySQL 5.6. An error now occurs. (Bug #24577194, Bug #82830) * Using SET PERSIST with the global_log system variable was ineffective. (Bug #24569624, Bug #82807) * In the Performance Schema variables_info table, the VARIABLE_SOURCE column was not set properly for some boolean options specified in option files without a variable. (Bug #24567960) * The .mylogin.cnf option file is intended for use by client programs, but the server was reading it as well. The server no longer reads it. (Bug #24557925) * Use of boost::chrono was replaced with std::chrono because the former was causing link errors. (Bug #24556808, Bug #82781) * The X Plugin was built with compilation options different from other plugins. (Bug #24555770, Bug #82777) * When populating the variables_by_thread table, the Performance Schema could attempt to access session variables of other threads that were being deinitialized. (Bug #24555658) * Users without proper privileges could load and unload server components. (Bug #24528148) * Concurrent execution of INSTALL COMPONENT and UNINSTALL COMPONENT statements could cause a server exit. (Bug #24527148) * On Debian/Ubuntu platforms, the systemd startup script for MySQL ignored datadir settings in /etc/mysql/my.cnf. (Bug #24517024, Bug #82709) * With a LOCK TABLES statement in effect, DML statements on a table that had triggers could cause a server exit. (Bug #24506766) * When the WITH_INNODB_MEMCACHED CMake option is enabled, memcached now can be built on Solaris 11 and 12 using the Developer Studio 12.5 compiler. (Bug #24504155, Bug #82692) * Parallel slave threads running account-management statements could fail due to a race condition in handling privilege cache locks. (Bug #24503606) * Executing ALTER TABLE on a table that has triggers concurrently with other DDL operations could result in a corrupted data dictionary. (Bug #24497803) * If mysqladmin shutdown encountered an error determining the server process ID file, it displayed an error message that did not clearly indicate the error was nonfatal. It now indicates that execution continues. (Bug #24496214) * For the null_audit plugin, setting the null_audit_event_record system variable improperly could cause a server exit. This variable should be set only from within the null_audit plugin, so it is now read only. (Bug #24493829, Bug #82670) * In certain contexts, an expression such as IF(col_name > 5000, (1 / col_name), 5000) could get a type of DECIMAL(6,4), which would truncate results. (Bug #24492965, Bug #82668) * The data structure used for ZEROFILL columns could experience memory corruption, leading eventually to a server exit. (Bug #24489302) * Library search path handling on Solaris was incorrect. (Bug #24487934, Bug #82646) * Operation of the mysql-multi.server.sh script was based on my.cnf in the data directory. That option file is no longer used, so mysql-multi.server.sh has been removed. (Bug #24487870) * A query could produce incorrect results if MIN() or MAX() in a subquery referred to an indexed column. (Bug #24484060, Bug #82638, Bug #24657798, Bug #82965) * SHOW TRIGGERS output order could differ on Linux and Windows. (Bug #24482919, Bug #82637) * mysqld_safe attempted to read my.cnf in the data directory, although that is no longer a standard option file location. (Bug #24482156) * After running mysql_upgrade, executing an INSTALL COMPONENT statement could cause a server exit. (Bug #24453571) * A regular expression pattern match into a large string could result in a server exit due to memory allocation failure or integer overflow. (Bug #24449076, Bug #24449090) * Starting the server with a MySQL 5.7 data directory resulted in failure due to absence of the default_roles and role_edges system tables. (Bug #24447771) * An incorrect error was reported for CREATE TABLE statements with a large value for the CONNECTION table option. The value is now limited to 1024 bytes. (Bug #24437124) * An assertion could be raised if an ER_LOCK_WAIT_TIMEOUT error occurred during execution of SHOW CREATE TRIGGER. (Bug #24420809, Bug #82483) * Constant folding could produce incorrect results for large unsigned integers. (Bug #24401273, Bug #82425) * Use of very long subpartition names could result in a server exit. Now partition or subpartition names larger than 64 characters produce an ER_TOO_LONG_IDENT error. (Bug #24400628, Bug #82429) * The Gis_wkb_vector<Gis_point> copy constructor was not explicitly instantiated, causing build problems for the Intel compiler. (Bug #24397833, Bug #82358) * Privilege escalation was possible by exploiting the way REPAIR TABLE used temporary files. (Bug #24388746) * A race condition between UNINSTALL PLUGIN and SHOW PLUGINS could result in a server exit. (Bug #24344026) * With ROW mode binary logging, component installation raised an assertion. (Bug #24343582) * A potential memory leak related to roles was fixed. (Bug #24337928) * Subqueries that were converted to semi-joins and programmatically generated an INFORMATION_SCHEMA table could incorrectly treat the INFORMATION_SCHEMA table as empty. A workaround for this problem prior to the bug fix: SET optimizer_switch='semijoin=off'; (Bug #24287772, Bug #82214) * mysqldumpslow failed to parse timestamps in the slow query log; it had not been updated to track a change in log timestamp format. (Bug #24007040) * The AppArmor profile installed by Ubuntu packages was missing an entry permitting libnuma to read a /sys hierarchy path, resulting in server startup failure. (Bug #23854929) * The optimizer could incorrectly treat RAND() as a constant for queries subjected to semi-join transformation. (Bug #23854015) * For debug builds, unequal-length binary operands for bit operators could raise an assertion. (Bug #23853628) * If a transaction rollback request rolled back only a statement and not the whole transaction, any attempt at attaching another operation to the transaction raised an assertion. (Bug #23753319, Bug #82143) * STRCMP() for arguments with the utf8mb4_unicode_ci collation could return results different from ORDER BY. STRCMP() has been corrected. (Bug #23752284, Bug #82132) * Complete logical backups made with mysqlpump could not be restored if GTIDs were enabled. (Bug #23748432) * Infinite recursion could occur if the audit_log plugin signalled an error while handling an error. (Bug #23717558, Bug #82052) * Messages written by the audit_log plugin to the error log regarding MYSQL_AUDIT_CONNECT event failures now print the underlying error cause as well to aid debugging. (Bug #23710632) * MySQL now uses readdir() rather than readdir_r(). The latter has been deprecated since glibc 2.24 and caused debug builds of MySQL and builds using GCC 6.1 to fail. Additionally, several problems resulting in GCC 6.1 compiler warnings were corrected. (Bug #23708395, Bug #24437737, Bug #82515, Bug #24459890, Bug #25103242) * FORCE INDEX was ineffective for SELECT COUNT(*) queries. (Bug #23596760, Bug #81854) * During startup, the server creates a lock file for the Unix socket file (for example, mysql.sock.lock as a lock file for mysql.sock). If the server failed to write the process ID to the lock file, it failed to remove that file, which could cause subsequent server startups to fail until the file was removed manually. (Bug #23582603, Bug #81838) * For debug builds, queries executed using Item_func_spatial_collection::val_str() could raise an assertion. (Bug #23573720) * For audit log events in the connection class, the connection_type value was available only for connect events. The value is now available in connect, disconnect, and change-user events. (Bug #23541550) * The audit_log plugin audit_log_filter_remove_filter() function caused a server exit if given a NULL argument. (Bug #23522793) * Attempts to configure MySQL without the FEDERATED storage engine failed (the CMake -DWITH_FEDERATED_STORAGE_ENGINE=0 option did not work). (Bug #23508203, Bug #81665) * On Solaris, gettimeofday() could return an invalid value and cause a server shutdown. (Bug #23499695) * The keyring_file plugin could attempt to write keys to its storage file when the file did not exist. To ensure that keys are flushed only when the correct storage file exists, keyring_file now stores a SHA-256 checksum of the keyring in the file. Before updating the file, the plugin verifies that it contains the expected checksum. (Bug #23498254) * START GROUP REPLICATION uses stacked Srv_session and did not return to the correct thread. START GROUP REPLICATION and STOP GROUP REPLICATION are now removed from the list of permitted commands. (Bug #23337984) * A union query resulting in tuples larger than max_join_size could result in a server exit. (Bug #23303485) * Grant tables with incorrect structure may cause problems in user management operations. As a consequence of the fix for this, for any operation that modifies a grant table, the server now checks whether the table has the expected structure and produces an error if not. mysql_upgrade must be run to update the tables to the expected structure. (Bug #23295423, Bug #25095876, Bug #25448037) * ST_ExteriorRing() could cause a server exit due to being passed an invalid WKB string believed to be valid. (Bug #23280574) * For debug builds, failure to prepare a branch of an XA transaction could lead to a server exit. (Bug #23264552, Bug #81375) * The optimizer could choose ref access on a secondary index rather than range access on the primary key, even when the cost was higher. (Bug #23259872, Bug #81341) * An instance of the disk-full error message contained the wrong error code. (Bug #23247332, Bug #81346) * For a query with ORDER BY and LIMIT, an optimizer trace did not record the optimizer's switch to a different index. (Bug #23227428, Bug #81250) * Improper handling of a lock used by the version_tokens plugin and user-defined functions could result in a server exit if a UDF was called while version_tokens was being uninstalled. (Bug #23210850) * The QUOTE() function could allocate excessive memory. A limit of max_allowed_packet bytes is now imposed and returns NULL with a warning for attempts to allocate more. (Bug #23195404) * For some deeply nested expressions, the optimizer failed to detect stack overflow, resulting in a server exit. (Bug #23135667) * For sessions created through the X Plugin, incorrect thread attachment/detachment could cause a server exit. (Bug #23057045) * When attempting to locate the data directory, mysqld_safe incorrectly considered $MY_BASEDIR_VERSION/var as one of the possible locations. (Bug #23013510, Bug #80866) * For some generated columns, character set conversion of the column definition for table rebuilds could change column values. (Bug #22991924) * If a query performed a GROUP BY on a column of a derived table and the select list contained an expression mixing an aggregate function and the group column, an error was raised if the ONLY_FULL_GROUP_BY SQL mode was enabled. (Bug #22924183, Bug #80726) * A binary (in-place) upgrade from MySQL 5.6 to 5.7 followed by a data export performed using mysqlpump resulted in an Invalid default value for date_column error for attempts to reload the dump file. (Bug #22919028, Bug #80706) * A failed cast of a long float to integer could cause a server exit. (Bug #22907691) * On Unix and Unix-like systems, the error log file resulting from specifying --log-error without an option value was incorrectly created in the directory of the PID file if the --pid-file option was also given. (Bug #22900354) * The main.log_tables-big test case could be unstable on highly loaded hosts. Thanks to Laurynas Biveinis for the patch. (Bug #22874167, Bug #80607) * SQL statements executed through the X Plugin were not instrumented in the Performance Schema. (Bug #22859462) * The rpl.rpl_key_rotation test case did not synchronize properly with the master server. Thanks to Laurynas Biveinis for the patch. (Bug #22838596, Bug #80531) * DROP INDEX operations could fail due to inconsistent handling of index prefix lengths for TEXT-type columns (TINYTEXT and so forth). (Bug #22740093, Bug #80392) * Metadata locking on stored routine names was performed on a case-sensitive basis, but routine names are not case sensitive. (Bug #22700385) * Queries that used an aggregate function with DISTINCT could produce incorrect results. (Bug #22686994, Bug #80310) * The innodb_numa_interleave system variable was erroneously available on some systems that were not NUMA-enabled. Thanks to Tomislav Plavcic for the patch. CMake now sets the default WITH_NUMA value based on whether the current platform has NUMA support. For platforms without NUMA support, CMake behaves as follows: + With no NUMA option (the normal case), CMake continues normally, producing only this warning: NUMA library missing or required version not available + With -DWITH_NUMA=ON, CMake aborts with this error: NUMA library missing or required version not available (Bug #22678436, Bug #80288) * Certain SQL queries involving complex WHERE conditions could cause warnings, memory corruption, or a server exit. (Bug #22671573) * When taking the server offline, a race condition within the Performance Schema could lead to a server exit. (Bug #22551677) * On macOS, if a table with an associated trigger was renamed to a new name containing both lowercase and uppercase characters, DROP TRIGGER for the trigger resulted in an ER_NO_SUCH_TABLE error for the table. (Bug #22512899, Bug #79873) * In the MYSQL_FIELD C API structure, the org_table value for derived tables was *, which could cause failure for queries that depend on this value. The org_table value for views and derived tables now is set as follows: If the column is selected from a view, org_table names the view. If the column is selected from a derived table, org_table names the base table. If a derived table wraps a view, org_table still names the base table. If the column is an expression, org_table is the empty string. (Bug #22364401, Bug #79641) * For RPM packages, the default error-log location in the deployed /etc/my.cnf file differed from the location in the installed logrotate script, causing logrotate to fail. (Bug #22322685) * The Performance Schema events_statements_summary_by_digest table could contain multiple rows for the same statement digest and schema combination, rather than the expected single (unique) row. (Bug #22320066, Bug #79533) * For Performance Schema system and status variable tables, variable values expressed in a character set different from utf8 could be truncated or incorrect. (Bug #22313205) * For debug builds, altering a table partitioning expression using an expression attribute with an invalid UTF-8 name caused an assertion to be raised. (Bug #22152229) * After performing inserts in a table containing an AUTO_INCREMENT column and then performing a SELECT operation, the LAST_INSERT_ID() returns the correct value, but the value of the mysql_insert_id() C API function was being reset to 0. (Bug #22028117, Bug #78778) * On Ubuntu, error messages were displayed during upgrades from Community to Commercial packages that made it appear as though mysqld and my_print_defaults had not been installed. Those messages were spurious and have been silenced. (Bug #21807248) * With the use_index_extensions flag of the optimizer_switch system variable disabled, some SELECT DISTINCT queries could return incorrect results. (Bug #21749123, Bug #78244) * An invalid string value in the WHERE clause of an UPDATE statement, caused an index scan rather than a range scan to be used. For values not present in the index, this could be much slower. Now the optimizer determines this to be an "impossible WHERE" condition. (Bug #21032418, Bug #76933) * Debian packages were missing an AppArmor-related include file and incorrectly were marked dependent on AppArmor (making it impossible to disable AppArmor by uninstalling it). (Bug #20768958) * When the automatic_sp_privileges system variable was enabled, it did not have the expected effect for anonymous users. (Bug #20266641) * In a replication environment, SET PASSWORD or ALTER USER could fail to execute on the slave due to failure to parse the hash string correctly. (Bug #20228478) * The optimizer resolve_const_item() function called Item_decimal() with the last two arguments in the wrong order. (Bug #19062566) * An in-place ALTER TABLE operation failed to report an error when adding a DATE or DATETIME column under these conditions: a) the column was NOT NULL and no default value was supplied; b) strict and NO_ZERO_DATE SQL modes were enabled; c) the table was not empty. An ALTER TABLE operation failed with an error rather than a warning when adding a DATE or DATETIME column under these conditions: a) the column was NOT NULL and no default value was supplied; b) strict SQL mode was enabled and NO_ZERO_DATE SQL mode was not enabled; c) the table was not empty. (Bug #16888677) * Inserting a TIME, DATE, or TIMESTAMP value with a fractional seconds part into a column having the same type but fewer fractional digits resulted in rounding. This differs from MySQL 5.5, which used truncation rather than rounding. To enable control over this behavior, a new TIME_TRUNCATE_FRACTIONAL SQL mode is available. The default is to use rounding. If this mode is enabled, truncation occurs instead. (Bug #16583910, Bug #68760) * NOT IN subqueries could produce incorrect results when an index prefix of an inner table was used. (Bug #13915291) * On non-Linux Unix systems, the mysql.server startup script used the Linux command pidof rather than pgrep. (Bug #13788154, Bug #64342) * Starting multiple instances of mysqld_safe after an abnormal server exit could result in one mysqld_safe instance killing another. As a consequence of the bug fix, the mysqld_safe.pid file is no longer used. (Bug #11751149, Bug #41908) * The --help message for mysqld_safe was corrected to mention that the --no-defaults, --defaults-file, and --defaults-extra-file options, if given, must be the first argument. (Bug #11745176, Bug #11192) * The bounds check for the XML parser position stack for each level (which has a fixed depth) used the size of the array as the upper limit, and so was off by one. This is fixed by decreasing the allowable depth by one, which actually matches the maximum number of elements in the position stack. (Bug #83871, Bug #25111907) References: See also: Bug #14040071, Bug #15948580. -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/mysql