|
------------------------------------------------------------
revno: 4805 committer: Marc Alff <marc.alff@oracle.com> branch nick: mysql-5.6.10-release-backport timestamp: Tue 2013-01-22 17:34:23 +0100 message: Bug#16196158 - RACE CONDITION IN TABLE PERFORMANCE_SCHEMA.THREADS Manual backport from mysql-5.6 (5.6.11) to mysql-5.6.10-release (5.6.10) ------------------------------------------------------------ revno: 4804 committer: Georgi Kodinov <georgi.kodinov@oracle.com> branch nick: mysql-5.6.10-release timestamp: Tue 2013-01-15 17:37:35 +0200 message: merge of bug #16103348 to the 5.6.10 release. ------------------------------------------------------------ revno: 4803 committer: Balasubramanian Kandasamy <balasubramanian.kandasamy@oracle.com> branch nick: mysql-5.6.10-release timestamp: Fri 2013-01-11 11:23:04 +0100 message: Removed Conflicts: mysql-libs mysql-libs-advanced from spec file to fix upgrade fix ------------------------------------------------------------ revno: 4802 committer: Marko M?kel? <marko.makela@oracle.com> branch nick: mysql-5.6.10-release timestamp: Thu 2013-01-10 11:48:56 +0200 message: Bug#16095573 Table is lost during 'ALTER TABLE' command (5.6.10 work-around) In 5.6.10, we prohibit DROP FOREIGN KEY with any of the operations that involve rebuilding the table (WL#6255). The real fix is in Bug#15989081 INNODB ALTER TABLE IS NOT ATOMIC. ------------------------------------------------------------ revno: 4801 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Tue 2013-01-08 13:59:16 +0100 message: Need to repush ------------------------------------------------------------ revno: 4800 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Tue 2013-01-08 12:54:37 +0100 message: Merge 16038656 follow-up from 5.6 ------------------------------------------------------------ revno: 4799 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Tue 2013-01-08 12:52:19 +0100 message: Bug#16038656 - INNODB_FT_USER_STOPWORD_TABLE BECOMES CORRUPT, CAUSES SERVER CRASH Bug#16051857 : REPLICATION FAIL TO START AFTER PROVISIONING SERVER FROM BACK DATADIR WITH GTIDS Bug#16066243 PB2 FAILURES I_MAIN.BUG15912213 AND I_MAIN.CTYPE_UTF8 FOR MACOSX10.6 FOR 5.1 Merged from 5.6 to 5.6.10-release ------------------------------------------------------------ revno: 4798 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Tue 2013-01-08 12:50:23 +0100 message: Bug#16036699 - FTS PARSER CANNOT HANDLE DASH IN TABLE NAME == CRASHES Bug#16046140 BIN/MYSQLD_SAFE: TEST: ARGUMENT EXPECTED Merged from 5.6 to 5.6.10-release ------------------------------------------------------------ revno: 4797 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Fri 2012-12-21 13:36:04 +0100 message: Fix mysql_config_editor test after new copyright year ------------------------------------------------------------ revno: 4796 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Fri 2012-12-21 13:07:46 +0100 message: Followup on collections cmake: added to .bzrignore ------------------------------------------------------------ revno: 4795 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Thu 2012-12-20 12:34:11 +0100 message: Update copyright year ------------------------------------------------------------ revno: 4794 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Thu 2012-12-20 12:31:35 +0100 message: Remove some internal comments from mysql_install_db ------------------------------------------------------------ revno: 4793 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-5610 timestamp: Thu 2012-12-20 12:27:26 +0100 message: A bit more intelligent processing of .in files in mysql-test/collections ------------------------------------------------------------ revno: 4792 tags: clone-5.6.10-build committer: Nuno Carvalho <nuno.carvalho@oracle.com> branch nick: mysql-5.6-bf timestamp: Fri 2012-12-14 15:31:30 +0000 message: BUG#15956714: REPLICATION DOESN'T WORK IN MYSQL 5.6.80-RC IF STARTED WITH SKIP-INNODB When a slave was started with --skip-innodb and info file repositories (default values), replication was incorrectly stopped. If slave is not doing a migration between info repositories and is using file repositories, replication can work flawless. When server is started with info file repositories and it is not doing a live migration, ignore errors when trying to open table info repositories. ------------------------------------------------------------ revno: 4791 [merge] committer: Luis Soares <luis.soares@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 15:18:40 +0000 message: BUG#15992220 Automerged bug branch into latest 5.6. ------------------------------------------------------------ revno: 4786.1.6 committer: Luis Soares <luis.soares@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 15:02:07 +0000 message: BUG#15992220 Result update for one test case (missed the additional field). ------------------------------------------------------------ revno: 4786.1.5 committer: Luis Soares <luis.soares@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 13:09:28 +0000 message: BUG#15992220 More tests cleanup. And added an include/have_gtid.inc to the test introduced by this patch. ------------------------------------------------------------ revno: 4786.1.4 committer: Luis Soares <luis.soares@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 12:48:33 +0000 message: BUG#15992220 Added a simple test showing that the new column on SHOW SLAVE STATUS, actually shows a different value everytime the CHANGE MASTER TO command changes AUTO_POSITION. ------------------------------------------------------------ revno: 4786.1.3 committer: Luis Soares <luis.soares@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 12:34:28 +0000 message: BUG#15992220 Addressing review comments. (Test cleanup.) ------------------------------------------------------------ revno: 4786.1.2 committer: Luis Soares <luis.soares@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 12:08:11 +0000 message: BUG#15992220: MAKE IT POSSIBLE TO DETERMINE IF GTID AUTO-POSITIONING PROTOCOL IS IN USE When GTIDs were introduced, a new master-slave protocol with auto-positioning was also introduced. The new protocol is turned on by: "CHANGE MASTER TO MASTER_AUTO_POSITION = 1" However, but there is currently no way to determine if the protocol is used or not. This patch extends the fields in SHOW SLAVE STATUS with an additional field: Auto_Position = (0|1), which states whether the new protocol is in use (1) or not (0). ------------------------------------------------------------ revno: 4790 [merge] committer: Ramil Kalimullin <ramil.kalimullin@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 18:44:39 +0400 message: Auto-merge. ------------------------------------------------------------ revno: 4788.1.2 committer: Ashish Agarwal<ashish.y.agarwal@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 19:41:47 +0530 message: Bug#16002890: FIX ERROR LOGGING FOR PLUGINS PROBLEM: my_plugin_log_message() takes a void** argument, but treats it as a void*.The result is undefined behaviour. SOLUTION: It takes void** argument and treats it as void**. ------------------------------------------------------------ revno: 4788.1.1 committer: Pedro Gomes <pedro.gomes@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 12:51:29 +0000 message: Bug#15985893: GTID SKIPPED EVENTS ON MASTER CAUSE SEMI SYNC TIME-OUTS hen using GTIDs and semi sync, whenever a transaction is executed on the master but is not sent to the slave because it already has this id, semi sync will time out. In fact, semi sync hooks will still trigger a waiting process in the master, but as the transaction never reaches the slave, a reply is never sent. Making the semi sync plug-in call the reportReplyBinlog method whenever a transaction is skipped solves this issue. @semisync_master.cc A new method skipSlaveReply was created to allow the interruption of the waiting process without a reply. @semisync_master.h New method header and other headers updated @semisync_master_plugin.cc The after send event response now depends on the transaction being skipped or not @replication.h rpl_handler.h rpl_handler.cc The after_send_event hook now takes new parameters that flag if the event was skipped, what was is position and binlog. @rpl_master.cc The after_send_event hook is now always invoked taking as new parameters the position and binlog when transactions are skipped. ------------------------------------------------------------ revno: 4789 committer: Ramil Kalimullin <ramil.kalimullin@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 17:53:26 +0400 message: Fix for bug #14503584. The library can't currently handle properly calculations involving self-intersecting polygons. We get an assertion when we try. Until proper support for that is implemented we will replace the assertion with an execution error. ------------------------------------------------------------ revno: 4788 [merge] committer: Ramil Kalimullin <ramil.kalimullin@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 15:23:52 +0400 message: Auto-merge. ------------------------------------------------------------ revno: 4786.1.1 committer: Jon Olav Hauglid <jon.hauglid@oracle.com> branch nick: mysql-5.6-bug15996571 timestamp: Fri 2012-12-14 11:16:24 +0100 message: Bug#15996571: WARNING RECOMMENDS TO USE A DEPRECATED OR NON-EXISTING OPTION The problem was that if the startup options --log-slow-admin-statements, --log-queries-not-using-indexes or --log-slow-slave-statements were used without specifying --slow-query-log, a warning message referring to the removed "--log_slow_queries" would be given instead of using the current name "--slow-query-log". This patch fixes the problem by updating the warning message. The patch also fixes similar problems with warning messages given if --general-log-file or --slow-query-log-file are used with --log-output=table (and not =file). ------------------------------------------------------------ revno: 4787 [merge] committer: Ramil Kalimullin <ramil.kalimullin@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 14:30:04 +0400 message: Manual merge from mysql-5.5. ------------------------------------------------------------ revno: 2875.417.95 [merge] committer: Ramil Kalimullin <ramil.kalimullin@oracle.com> branch nick: mysql-5.5 timestamp: Fri 2012-12-14 14:01:43 +0400 message: Auto-merge from mysql-5.1. ------------------------------------------------------------ revno: 2661.862.50 committer: Ramil Kalimullin <ramil.kalimullin@oracle.com> branch nick: mysql-5.1 timestamp: Fri 2012-12-14 13:55:30 +0400 message: Fix for BUG#15948580 UPDATE_XML() CRASHES THE SERVER. Problem: tag's buffer overflow leads to a problem. Fix: bound check added. ------------------------------------------------------------ revno: 4786 committer: Bill Qu <bill.qu@Oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 15:09:26 +0800 message: Approved by Jimmy and Inaam. rb#1576 ------------------------------------------------------------ revno: 4785 committer: Jimmy Yang <jimmy.yang@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 14:47:54 +0800 message: Fix Bug #15994393 - VALGRIND FAILURE ON MYSQL-5.6 IN FTS SHUTDOWN / FTS_OPTIMIZE_THREAD rb://1717 approved by Sunny Bains ------------------------------------------------------------ revno: 4784 [merge] committer: Inaam Rana <inaam.rana@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 11:35:34 +0500 message: merge from 5.5 ------------------------------------------------------------ revno: 2875.417.94 [merge] committer: Inaam Rana <inaam.rana@oracle.com> branch nick: mysql-5.5 timestamp: Fri 2012-12-14 11:29:07 +0500 message: merge from 5.1 ------------------------------------------------------------ revno: 2661.862.49 committer: Inaam Rana <inaam.rana@oracle.com> branch nick: mysql-5.1 timestamp: Fri 2012-12-14 11:24:57 +0500 message: Bug#14329288 IS THE CALL TO IBUF_MERGE_OR_DELETE_FOR_PAGE FROM BUF_PAGE_GET_GEN REDUNDANT? rb://1711 approved by: Marko Makela When decompressing a compressed page that had already been accessed in the buffer pool, do not attempt to merge buffered changes. ------------------------------------------------------------ revno: 4783 committer: Inaam Rana <inaam.rana@oracle.com> branch nick: mysql-5.6 timestamp: Fri 2012-12-14 10:46:05 +0500 message: Bug#14774779 CRASH RECOVERY FAILED, UNABLE TO COMPRESS A PAGE rb://1721 approved by: Marko Makela The bug was introduced by WL6344 in the innodb_log_compressed_pages part. If we don't log compressed pages then we need to do compression during recovery which requires that the uncompressed pages must be deterministically same during recovery but they can differ in garbage bytes potentially causing compression during recovery to fail. This bug will manifest itself only when recovery is done in multiple batches of redo apply. This workaround is to take away innodb_log_compressed_pages option. Note it is not a full revert of WL6344 ------------------------------------------------------------ revno: 4782 committer: Mattias Jonsson <mattias.jonsson@oracle.com> branch nick: b15961327-56 timestamp: Thu 2012-12-13 13:38:18 +0100 message: Bug#15961327: ALTER TABLE...MODIFY...NOT NULL REFUSES TO SILENTLY CONVERT NULL In-Place in InnoDB do not convert from NULL to NOT NULL due to loss of information, it will always give an error for such rows. But ALGORITHM = COPY will do such conversion with only a warning (which in case of strict sql_mode it will turn the warning to an error). So to be consistent, we will only allow INPLACE for NULL -> NOT NULL under strict sql_mode, since that will give an error for both INPLACE and COPY. If not strict mode ha_innodbase::check_if_supported_inplace_alter will return failure, and depending on ALGORITHM it will fall back to COPY or fail. Also notice that ADD PRIMARY KEY () will implicitly add NOT NULL if not already there! Updated tests that uses INPLACE by setting strict sql_mode. Also updated some Swedish error messages. rb 1679 approved by marko. ------------------------------------------------------------ revno: 4781 [merge] committer: Ravinder Thakur <ravinder.thakur@oracle.com> branch nick: 5.6 timestamp: Thu 2012-12-13 22:36:33 +0530 message: Null merge from 5.5 to 5.6 for bug#11761752 ------------------------------------------------------------ revno: 2875.417.93 [merge] committer: Ravinder Thakur <ravinder.thakur@oracle.com> branch nick: 5.5 timestamp: Thu 2012-12-13 20:58:09 +0530 message: Merging from 5.1 to 5.5 for bug#11761752 ------------------------------------------------------------ revno: 2661.862.48 committer: Ravinder Thakur <ravinder.thakur@oracle.com> branch nick: 5.1 timestamp: Thu 2012-12-13 20:33:44 +0530 message: bug#11761752: DO NOT ALLOW USE OF ALTERNATE DATA STREAMS ON NTFS FILESYSTEM. File names with colon are being disallowed because of the Alternate Data Stream (ADS) feature of NTFS that could be misused. ADS allows data to be written to alternate streams of a normal file. The data in alternate streams cannot be seen by normal tools on Windows (explorer, cmd.exe). As a result someone can use this feature to hide large amount of data in alternate streams and admins will have no easy way of figuring out the files that are using that disk space. The fix also disallows ADS in the scenarios where file name is passed as some dynamic variable. An important thing about the fix is that it DOES NOT disallow ADS file names if they are not dynamic (i.e. if the file is created by using some option that needs local access to the MySQL server, for example error log file). The reasoning is that if some MySQL option related to files requires access to the local machine (it is not dynamic), then user can very well create data in ADS by some other means. This fixes only those scenarios which can allow users to create data in ADS over the wire. File names with colon are being disallowed only on Windows. UNIX (Linux in particular) supports NTFS, but it will not be a common scenario for someone to configure a NTFS file system to store MySQL data on Linux. Changes in file bug11761752-master.opt are needed due to bug number 15937938. ------------------------------------------------------------ revno: 4780 committer: Ritheesh Vedire <ritheesh.vedire@oracle.com> branch nick: Bug14775893 timestamp: Thu 2012-12-13 22:16:57 +0530 message: Bug#14775893: MASTER_POS_WAIT CAN HANG WHEN USING GTIDS Bug#15927032: SEMI SYNC REPLICATION BREAKS WITH CHANGE MASTER_AUTO_POSITION=1 When transactions are skipped in a GTID protocol, the coordinates of the binary log (on the master) are not updated properly by the Slave IO (mi) and Slave SQL (rli) threads. This can have several consequences. For example, master_pos_wait() depends on the coordinates of the SQL thread, and the function can hang forever or return -1 due to improper updation of rli. Other example is, during semi sync replication, the master receives an ACK with wrong coordinates by slave IO which will break semi sync repl. This patch corrects the coordinates of the slave by sending a heart beat event which carries the coordinates of the last skipped transaction during a gtid protocol. This event when received by the IO thread, is put in to the relay log as an ignored Rotate Event. The Rotate Event is read by the slave SQL thread to update rli. Thereby, this patch solves following problems 0) Fixes the *holes* in event coordianates of mi and rli 1) master_pos_wait() returns values >=0 2) Fixes the problems in semi sync when using with GTIDs ------------------------------------------------------------ revno: 4779 committer: Manish Kumar<manish.4.kumar@oracle.com> branch nick: mysql-5.6_15907706 timestamp: Thu 2012-12-13 21:55:08 +0530 message: BUG#15907706 : RPL.RPL_INSERT_ID TEST FAILS ON SOLARIS PLATFORM IN PB2 Problem : rpl.rpl_insert_id fails on PB2 everyday and only on solaris platform. The problem was that the thread_stack value available for the test seems to be less when the test executes on the solaris machines. On other platforms the default value 196608 for 32 bit machines and 262144 bytes for 64 big machines seems to be enough. But on linux the error reports that the the thread_stack value required is beyond the specified limit on the respective platforms. Fix - The problem is fixed by passing the PB_HOST_SPECIFIC_MYSQLD_ARGS host specific arguments to the slave. ------------------------------------------------------------ revno: 4778 committer: Ravinder Thakur <ravinder.thakur@oracle.com> branch nick: 5.6 timestamp: Thu 2012-12-13 21:04:59 +0530 message: Merging from 5.5 to 5.6 for bug#11761752 In addition to changes from 5.5, the fix for 5.6 contains additional checks for InnoDB option innodb_buffer_pool_filename. This option does not exists in 5.5 or 5.1. ------------------------------------------------------------ revno: 4777 [merge] committer: Marko M?kel? <marko.makela@oracle.com> branch nick: mysql-5.6 timestamp: Thu 2012-12-13 17:14:56 +0200 message: Merge mysql-5.5 to mysql-5.6. ------------------------------------------------------------ revno: 2875.417.92 committer: Marko M?kel? <marko.makela@oracle.com> branch nick: mysql-5.5 timestamp: Thu 2012-12-13 17:12:21 +0200 message: Follow-up fix to Bug#14628410: Remove the Windows InnoDB Plugin specific implementation of innobase_mysql_tmpfile() from MySQL 5.5 onwards. ------------------------------------------------------------ revno: 4776 committer: Libing Song <libing.song@oracle.com> branch nick: mysql-5.6 timestamp: Thu 2012-12-13 22:50:52 +0800 message: Bug#15980626 POSSIBLE REPLICATION PERFORMANCE DEGRADATION THROUGH RLI->FLUSH_INFO(TRUE) Problem:sync_relay_log_info was being ignored when transactional tables were updated. No matter what number was set, relay log info file was synced upon committing. This was causing the performance degradation. Fix: Sync(or update relay log info table) after each transaction should happen only when both data and relay log info are stored in transactional tables. Only this scenario will guarantee crash safety. ------------------------------------------------------------ revno: 4775 committer: Dmitry Lenev <Dmitry.Lenev@oracle.com> branch nick: mysql-5.6-mdl-t-fix timestamp: Thu 2012-12-13 17:59:02 +0400 message: Follow-up for the fix for bug #15954872 "MAKE MDL SUBSYSTEM AND TABLE DEFINITION CACHE ROBUST AGAINST BUGS IN CALLERS". Make mdl-t unit test pass on Solaris and Windows by adjusting patterns used in EXPECT_DEATH macros to take into account different messages about assertion failures on these platforms. ------------------------------------------------------------ revno: 4774 [merge] committer: Satya Bodapati <satya.bodapati@oracle.com> branch nick: mysql-5.6-bug-14628410 timestamp: Thu 2012-12-13 19:07:39 +0530 message: Merge fix for Bug#14628410 from mysql-5.5 to mysql-5.6 ------------------------------------------------------------ revno: 2875.417.91 [merge] committer: Satya Bodapati <satya.bodapati@oracle.com> branch nick: mysql-5.5-bug-14628410 timestamp: Thu 2012-12-13 18:56:47 +0530 message: Merge fix for Bug#14628410 from mysql-5.1 to mysql-5.5 ------------------------------------------------------------ revno: 2661.862.47 committer: Satya Bodapati <satya.bodapati@oracle.com> branch nick: mysql-5.1-bug-14628410 timestamp: Thu 2012-12-13 18:53:16 +0530 message: Bug#14628410 - ASSERTION `! IS_SET()' FAILED IN DIAGNOSTICS_AREA::SET_OK_STATUS The error code returned from Merge file/Temp file creation functions are ignored. Use the return codes of the row_merge_file_create() and innobase_mysql_tmpfile() to return the error to caller if file creation fails. Approved by Marko. rb#1618 ------------------------------------------------------------ revno: 4773 committer: Alexander Nozdrin <alexander.nozdrin@oracle.com> branch nick: 5.6 timestamp: Thu 2012-12-13 15:02:05 +0400 message: A patch for Bug#15877062: MYSQL_SQL_STMT_CLOSE READS FREED MEMORY. The problem was that the PS-arena was not properly restored in case of error. The fix is to restore query arena properly. ------------------------------------------------------------ revno: 4772 committer: Jimmy Yang <jimmy.yang@oracle.com> branch nick: mysql-5.6 timestamp: Thu 2012-12-13 18:23:15 +0800 message: Fix Bug #14771282 - FULLTEXT INDEX NOT BEING PROPERLY MAINTAINED ON INNODB TABLE rb://1698 approved by Sunny Bains ------------------------------------------------------------ revno: 4771 committer: Guilhem Bichot <guilhem.bichot@oracle.com> branch nick: 5.6-3 timestamp: Wed 2012-12-12 12:19:17 +0100 message: Bug#13897507 - EXPLAIN DOES NOT LIST SUBQUERIES IN EXECUTION ORDER this patch does not change the order of subqueries in EXPLAIN (too difficult to do it now). The execution order of subqueries (which subquery is executed first, etc.) can be determined by looking at "attached_condition" of the table to which this subquery is attached, in the output of EXPLAIN FORMAT =JSON. Because this condition is long to read, and there can be many '()', this patch adds, to any subquery printed in the condition, the number of this subquery (select#) (as is already the case in the warning of EXPLAIN EXTENDED and EXPLAIN FORMAT= JSON, and in the Optimizer trace). This number can then be matched with numbers of query_block's of subqueries. At line 813 of explain_json_none.result we can see that this numbering helps finding what subquery is executed first. ------------------------------------------------------------ revno: 4770 committer: Oystein Grovlen <oystein.grovlen@oracle.com> branch nick: mysql-5.6 timestamp: Thu 2012-12-13 09:22:59 +0100 message: Bug#15950531 - DBUG_ASSERT(FT_HANDLER); IN ITEM_FUNC_MATCH::GET_COUN If no full text index for the match expression, InnoDB will set error status during initialization, and not create any FTS handler. The new count(*) optimization hit an assert if there was no handler. Fix: Check for error status before calling Item_func_match::get_count(). ------------------------------------------------------------ revno: 4769 committer: Marko M?kel? <marko.makela@oracle.com> branch nick: mysql-5.6 timestamp: Thu 2012-12-13 09:46:15 +0200 message: Bug#14105491 ADD FOREIGN KEY ON COLUMNS BEING RENAMED DOES NOT WORK CORRECTLY Disable the combination of ADD_FOREIGN_KEY and ALTER_COLUMN_NAME for now. rb#1703 approved by Jimmy Yang ------------------------------------------------------------ revno: 4768 [merge] committer: Harin Vadodaria <harin.vadodaria@oracle.com> branch nick: 56_bug15965288 timestamp: Thu 2012-12-13 10:21:09 +0530 message: Bug#15965288: BUFFER OVERFLOW IN YASSL FUNCTION DOPROCESSREPLY() Description: Merge from 5.5 to 5.6 . ------------------------------------------------------------ revno: 2875.417.90 [merge] committer: Harin Vadodaria <harin.vadodaria@oracle.com> branch nick: 55_bug15965288 timestamp: Thu 2012-12-13 10:19:14 +0530 message: Bug#15965288: BUFFER OVERFLOW IN YASSL FUNCTION DOPROCESSREPLY() Description: Merge from 5.1 to 5.5 ------------------------------------------------------------ revno: 2661.862.46 committer: Harin Vadodaria <harin.vadodaria@oracle.com> branch nick: 51_bug15965288 timestamp: Thu 2012-12-13 10:17:26 +0530 message: Bug#15965288: BUFFER OVERFLOW IN YASSL FUNCTION DOPROCESSREPLY() Description: Function DoProcessReply() calls function decrypt_message() in a while loop without performing a check on available buffer space. This can cause buffer overflow and crash the server. This patch is fix provided by Sawtooth to resolve the issue. ------------------------------------------------------------ revno: 4767 committer: Sunny Bains <Sunny.Bains@Oracle.Com> branch nick: 5.6 timestamp: Thu 2012-12-13 13:12:11 +1100 message: Bug#4789249 - LOCKING ASSERT IN FK UPDATE CASCADE CODE, LOCK->TRX->LOCK.WAIT_LOCK == LOCK Bug#14707091 SEGV IN LOCK_DEADLOCK_LOCK_PRINT(), NULL LOCK PTR Joining T1 does a deadlock check due to wait T1 is selected as victim (but not rolled back), for any number of reasons. We try and resolve more deadlocks - T2 is selected as a victim and rolled back - This grants T1 the lock and is no longer the victim. Since T1 was originally selected as a victim, when we go to reset its lock state we trip over the invariant that it must still be in the wait state. Fix is to check if T1 was granted a lock when T2 was rolled back during additional deadlock checking. Approved by Jimmy Yang rb#1619. ------------------------------------------------------------ revno: 4766 [merge] committer: Nirbhay Choubey <nirbhay.choubey@oracle.com> branch nick: 5.6 timestamp: Wed 2012-12-12 22:44:50 +0530 message: Merge of patch for Bug#13639125 from mysql-5.5. ------------------------------------------------------------ revno: 2875.417.89 committer: Nirbhay Choubey <nirbhay.choubey@oracle.com> branch nick: 5.5 timestamp: Wed 2012-12-12 22:31:03 +0530 message: Bug#13639125 DELIMITER STRIPS THE NEXT NEW LINE IN A SQL STATEMENT While processing each lines entered at the prompt, mysql client appends a '\n' to all the lines except for delimiter commands. However the same logic must not apply if 'delimiter' is part of a string or a comment, for which a '\n' should be added. Fixed by adding appropriate checks. Added a test case. ------------------------------------------------------------ revno: 4765 committer: Sven Sandberg <sven.sandberg@oracle.com> branch nick: 5.6 timestamp: Wed 2012-12-12 17:07:02 +0100 message: BUG#15968413 - IMPROVE GTID_SET::IS_INTERSECTION AND GTID_SET::INTERSECTION PERFORMANCE Problem: 1. Gtid_set::is_intersection and Gtid_set::intersection were too slow, because they iterated over all GTIDs instead of over all intervals. So calling one of the functions on a Gtid_set with a single, very big interval like 1-1000000000000 would take too long time. 2. Gtid_set::is_subset had a bug: if this and super had different sid_maps, there was a missing bounds check on super's sidno. This could cause read from invalid memory and dereferencing uninitialized pointer if super had fewer sidnos allocated than super's sid_map, and this contains a sidno that super does not contain. Fix: 1. Iterate over intervals, not over GTIDs. 2. Move the bounds check so it's checked every time. Also: - Rename is_intersection to is_intersection_nonempty. - Refactor and document is_subset, just for clarity. - Simplify loops in is_subset. ------------------------------------------------------------ revno: 4764 [merge] committer: magnus.blaudd@oracle.com branch nick: 5.6 timestamp: Wed 2012-12-12 15:46:49 +0100 message: Merge ------------------------------------------------------------ revno: 4763.1.1 [merge] committer: magnus.blaudd@oracle.com branch nick: 5.6 timestamp: Wed 2012-12-12 15:40:58 +0100 message: Merge ------------------------------------------------------------ revno: 4627.1.1 committer: magnus.blaudd@oracle.com branch nick: 5.6 timestamp: Tue 2012-11-20 20:18:59 +0100 message: rpl_extra_col_master - fix problem with unpredictable result because of missing ORDER BY in new testcase ------------------------------------------------------------ revno: 4763 committer: Mayank Prasad <mayank.prasad@oracle.com> branch nick: 5.6 timestamp: Wed 2012-12-12 19:03:54 +0530 message: Bug#14494893 : INNODB_FILE_PER_TABLE DOES NOT PERMIT ON/OFF AT STARTUP Details : - During server initialization, when boolean values 1/true OR 0/false are given, then they were accpected but if on/off were given then they were being considered wrong. Reason : - During server initialization, while checking for the valid values for boolean parameters, only 1/true AND 0/false were being checked. Solution : - Added code to check for On/Off and if found, make them valid values. Test case innodb_file_per_table_basic.test is already there to check this parameter. So no new test case required. I just added --innodb_file_per_table in master.opt files with value On (Default) so that at startup this fix could be verified. ------------------------------------------------------------ revno: 4762 [merge] committer: Ashish Agarwal<ashish.y.agarwal@oracle.com> branch nick: bug_14843970 timestamp: Wed 2012-12-12 17:56:23 +0530 message: BUG#14843970: VALIDATE_PASSWORD ALLOWS BLANK PASSWORDS PROBLEM: Empty string password is accepted even if validate_password plugin is installed. SOLUTION: Not allowing empty string password when validate_password plugin is installed. ------------------------------------------------------------ revno: 4532.1.1 committer: Ashish Agarwal<ashish.y.agarwal@oracle.com> branch nick: bug14843970 timestamp: Thu 2012-11-08 17:39:33 +0530 message: BUG#14843970: VALIDATE_PASSWORD ALLOWS BLANK PASSWORDS PROBLEM: Empty string password is accepted even if validate_password plugin is installed. SOLUTION: Not allowing empty string password when validate_password plugin is installed. ------------------------------------------------------------ revno: 4761 committer: Georgi Kodinov <georgi.kodinov@oracle.com> branch nick: wl6587-5.6 timestamp: Wed 2012-12-12 13:41:12 +0200 message: Backport of WL#6587 to 5.6 : version 2 Proper unix file format used for the tests. ------------------------------------------------------------ revno: 4760 committer: Satya Bodapati <satya.bodapati@oracle.com> branch nick: mysql-5.6-bug-15913195 timestamp: Wed 2012-12-12 15:34:58 +0530 message: Bug#15913195 - FIL_MAKE_IBBACKUP_OLD_NAME FUNCTION CAN CAUSE BUFFER OVERFLOW Fixed the buffer overflow by passing the right offset of the buffer to ut_sprintf_timestamp_without_extra_chars() No testcase because this is fixed and tested in meb branches(mysql-5.5-meb and mysql-5.6-meb). It is possible only to test in meb branches because of the UNIV_HOTBACKUP flag Approved by Marko. rb#1620 ------------------------------------------------------------ revno: 4759 [merge] committer: Joerg Bruehe <joerg.bruehe@oracle.com> branch nick: mysql-5.6 timestamp: Wed 2012-12-12 10:54:23 +0100 message: Automerge ------------------------------------------------------------ revno: 4757.1.1 [merge] committer: sayantan.dutta@oracle.com branch nick: mysql-5.6 timestamp: Wed 2012-12-12 15:12:50 +0530 message: upmerge 14737171 5.5=>5.6 ------------------------------------------------------------ revno: 2875.417.88 [merge] committer: sayantan.dutta@oracle.com branch nick: mysql-5.5 timestamp: Wed 2012-12-12 15:10:47 +0530 message: upmerge 14737171 5.1=>5.5 ------------------------------------------------------------ revno: 2661.862.45 committer: sayantan.dutta@oracle.com branch nick: mysql-5.1 timestamp: Wed 2012-12-12 15:09:31 +0530 message: Bug #14737171:MTR DOES NOT PRESERVE TEST CASE LOGS ON RETRY-FAIL ------------------------------------------------------------ revno: 4758 [merge] committer: Joerg Bruehe <joerg.bruehe@oracle.com> branch nick: mysql-5.6 timestamp: Wed 2012-12-12 10:48:07 +0100 message: Merge 5.6.9 into main 5.6 ------------------------------------------------------------ revno: 4751.1.2 committer: Joerg Bruehe <joerg.bruehe@oracle.com> branch nick: merge-5.6 timestamp: Wed 2012-12-12 10:44:20 +0100 message: ULN RPM spec file: Keep in sync with traditional (names for generic). ------------------------------------------------------------ revno: 4751.1.1 [merge] committer: Joerg Bruehe <joerg.bruehe@oracle.com> branch nick: merge-5.6 timestamp: Tue 2012-12-11 14:19:16 +0100 message: Merge the 5.6.9 build back into main 5.6 ------------------------------------------------------------ revno: 4575.1.4 tags: mysql-5.6.9 committer: Bjorn Munch <bjorn.munch@oracle.com> branch nick: rel-569 timestamp: Thu 2012-11-22 15:12:33 +0100 message: Fix RPM name for generic linux, - not allowed |