Home | Back
------------------------------------------------------------
revno: 3260
tags: clone-5.5.20-build
committer: Sergey Vojtovich <sergey.vojtovich@oracle.com>
branch nick: mysql-5.5-security-push
timestamp: Fri 2011-12-16 12:17:13 +0400
message:
  Test case for BUG11763712 is intended for 5.1 only.
------------------------------------------------------------
revno: 3259 [merge]
committer: Mattias Jonsson <mattias.jonsson@oracle.com>
branch nick: topush-5.5-sec
timestamp: Thu 2011-12-15 17:12:04 +0100
message:
  merge. Disabled the added test in merge.test from 5.1.
    ------------------------------------------------------------
    revno: 2661.768.43
    tags: clone-5.1.61-build
    committer: Mattias Jonsson <mattias.jonsson@oracle.com>
    branch nick: topush-5.1-sec
    timestamp: Thu 2011-12-15 16:59:18 +0100
    message:
      Post push fix for merge.test and mysqlcheck.test on windows
------------------------------------------------------------
revno: 3258 [merge]
committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
branch nick: merge-5.5-security
timestamp: Thu 2011-12-15 14:17:21 +0200
message:
  empty weave merge mysql-5.1-security->mysql-5.5-security
    ------------------------------------------------------------
    revno: 2661.768.42 [merge]
    committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
    branch nick: merge-5.1-security
    timestamp: Thu 2011-12-15 14:10:20 +0200
    message:
      merge mysql-5.1->mysql-5.1-security
------------------------------------------------------------
revno: 3257 [merge]
committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
branch nick: merge-5.5-security
timestamp: Thu 2011-12-15 14:13:46 +0200
message:
  merge mysql-5.5->mysql-5.5-security
    ------------------------------------------------------------
    revno: 3097.116.70
    committer: Chaithra Gopalareddy<chaithra.gopalareddy@oracle.com>
    branch nick: mysql-5.5
    timestamp: Thu 2011-12-15 16:48:40 +0530
    message:
      Bug#13344643:Format function in view looses locale information
      
      Problem description:
      When a view is created using function FORMAT and if FORMAT function uses locale
      option,definition of view saved into server doesn't contain that locale information,
      Ex:
      create  table test2 (bb decimal (10,2));
      insert into test2 values (10.32),(10009.2),(12345678.21);
      create view test3 as select format(bb,1,'sk_SK') as cc from test2;
      select * from test3;
      +--------------+
      | cc           |
      +--------------+
      | 10.3         |
      | 10,009.2     |
      | 12,345,678.2 |
      +--------------+
      3 rows in set (0.02 sec)
      
      show create view test3
                      View: test3
               Create View: CREATE ALGORITHM=UNDEFINED DEFINER=`root`@`localhost`
      SQL SECURITY DEFINER VIEW `test3` AS select format(`test2`.`bb`,1) AS `cc`
      from `test2`
      character_set_client: latin1
      collation_connection: latin1_swedish_ci
      1 row in set (0.02 sec)
      
      Problem Analysis:
      The function Item_func_format::print() which prints the query string to create
      the view does not print the third argument (i.e the locale information). Hence
       view is created without locale information.
      
      Problem Solution:
      If argument count is more than 2 we now print the third argument onto the query string.
      
      Files changed:
      sql/item_strfunc.cc
      Function call changes: Item_func_format::print()
      mysql-test/t/select.test
      Added test case to test the bug
      mysql-test/r/select.result
      Result of the test case appended here
    ------------------------------------------------------------
    revno: 3097.116.69
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5
    timestamp: Thu 2011-12-15 10:44:33 +0100
    message:
      Bug#13463417 63487: ANNOYING TRACE MESSAGE IN CMAKE CODE
      
      Remove it.
    ------------------------------------------------------------
    revno: 3097.116.68
    committer: Andrei Elkin <andrei.elkin@oracle.com>
    branch nick: 5.5-bug13437900-repl_ignore_server_ids
    timestamp: Wed 2011-12-14 17:02:55 +0200
    message:
      
      bug#13437900
      post-push changes to please solaris compiler.
    ------------------------------------------------------------
    revno: 3097.116.67
    committer: Andrei Elkin <andrei.elkin@oracle.com>
    branch nick: 5.5-bug13437900-repl_ignore_server_ids
    timestamp: Wed 2011-12-14 15:33:43 +0200
    message:
      Bug#13437900 - VALGRIND REPORTS A LEAK FOR REPL_IGNORE_SERVER_IDS
      
      There was memory leak when running some tests on PB2.
      The reason of the failure is an early return from change_master()
      that was supposed to deallocate a dyn-array.
      
      Actually the same bug58915 was fixed in trunk with relocating the dyn-array
      destruction into THD::cleanup_after_query() which can't be bypassed.
      The current patch backports magne.mahre@oracle.com-20110203101306-q8auashb3d7icxho
      and adds two optimizations: were done: the static buffer for the dyn-array to base on,
      and the array initialization is called precisely when it's necessary rather than
      per each CHANGE-MASTER as before.
    ------------------------------------------------------------
    revno: 3097.116.66
    committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
    branch nick: B11754011-5.5
    timestamp: Wed 2011-12-14 12:49:32 +0200
    message:
      Addendum to the fix for bug #11754011: fixed a testcase result to include
      the new --slow-start-timeout option's help output
    ------------------------------------------------------------
    revno: 3097.116.65
    committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
    branch nick: B11754011-5.5
    timestamp: Tue 2011-12-13 17:44:19 +0200
    message:
      Bug#11754011: 45546: START WINDOWS SERVICE, THEN EXECUTE WHAT IS NEEDED.
      
      Added a global read-only option slow-start-timeout to control the
      Windows service control manager's service start timeout, that was
      currently hard-coded to be 15 seconds.
      The default of the new option is 15 seconds.
      The timeout can also be set to 0 (to mean no timeout applicable).
    ------------------------------------------------------------
    revno: 3097.116.64 [merge]
    committer: Annamalai Gurusami <annamalai.gurusami@oracle.com>
    branch nick: mysql-5.5-bug13117023
    timestamp: Tue 2011-12-13 17:10:09 +0530
    message:
      Merging from mysql-5.1 to mysql-5.5.
        ------------------------------------------------------------
        revno: 2661.798.2
        committer: Annamalai Gurusami <annamalai.gurusami@oracle.com>
        branch nick: mysql-5.1-bug13117023
        timestamp: Tue 2011-12-13 14:26:12 +0530
        message:
          Bug #13117023: Innodb increments handler_read_key when it should not
          
          The counter handler_read_key (SSV::ha_read_key_count) is incremented
          incorrectly.
          
          The mysql server maintains a per thread system_status_var (SSV)
          object.  This object contains among other things the counter
          SSV::ha_read_key_count. The purpose of this counter is to measure the
          number of requests to read a row based on a key (or the number of
          index lookups).
          
          This counter was wrongly incremented in the
          ha_innobase::innobase_get_index(). The fix removes
          this increment statement (for both innodb and innodb_plugin).
          
          The various callers of the innobase_get_index() was checked to
          determine if anybody must increment this counter (if they first call
          innobase_get_index() and then perform an index lookup).  It was found
          that no caller of innobase_get_index() needs to worry about the
          SSV::ha_read_key_count counter.
    ------------------------------------------------------------
    revno: 3097.116.63 [merge]
    committer: Marko M?kel? <marko.makela@oracle.com>
    branch nick: mysql-5.5
    timestamp: Mon 2011-12-12 13:53:18 +0200
    message:
      Merge mysql-5.1 to mysql-5.5.
        ------------------------------------------------------------
        revno: 2661.798.1
        committer: Marko M?kel? <marko.makela@oracle.com>
        branch nick: mysql-5.1
        timestamp: Mon 2011-12-12 13:48:24 +0200
        message:
          Bug#13418887 ERROR IN DIAGNOSTIC FUNCTION PAGE_REC_PRINT()
          
          When printing information about a ROW_FORMAT=REDUNDANT record, pass
          the correct flag to rec_get_next_offs().
          
          rb:821 approved by Jimmy Yang
------------------------------------------------------------
revno: 3256 [merge]
committer: Mattias Jonsson <mattias.jonsson@oracle.com>
branch nick: topush-5.5-sec
timestamp: Wed 2011-12-14 15:33:01 +0100
message:
  merged bug#12361113.
  
  Also added tests for partitions key caches.
    ------------------------------------------------------------
    revno: 2661.768.41 [merge]
    committer: Mattias Jonsson <mattias.jonsson@oracle.com>
    branch nick: topush-5.1-sec
    timestamp: Wed 2011-12-14 14:41:40 +0100
    message:
      merge
        ------------------------------------------------------------
        revno: 1810.4000.36 [merge]
        tags: clone-5.0.95-build
        committer: Mattias Jonsson <mattias.jonsson@oracle.com>
        branch nick: topush-5.0-sec
        timestamp: Wed 2011-12-14 14:05:22 +0100
        message:
          merge
            ------------------------------------------------------------
            revno: 1810.4005.1
            committer: Mattias Jonsson <mattias.jonsson@oracle.com>
            branch nick: b12361113-50-sec
            timestamp: Mon 2011-12-12 14:07:02 +0100
            message:
              Bug#12361113: CRASH WHEN "LOAD INDEX INTO CACHE" WITH TOO
              SMALL KEY CACHE
              
              The server crashed on division by zero because the key cache was not
              initialized and the block length was 0 which was used in a division.
              
              The fix was to not allow CACHE INDEX if the key cache was not initiallized.
              Thus never try LOAD INDEX INTO CACHE for an uninitialized key cache.
              
              Also added some windows files/directories to .bzrignore.
------------------------------------------------------------
revno: 3255 [merge]
committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
branch nick: merge-5.5-security
timestamp: Mon 2011-12-12 13:45:54 +0200
message:
  empty weave merge mysql-5.1-security->mysql-5-5.security
    ------------------------------------------------------------
    revno: 2661.768.40 [merge]
    committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
    branch nick: merge-5.1-security
    timestamp: Mon 2011-12-12 12:25:35 +0100
    message:
      merge 5.1->5.1-security.
------------------------------------------------------------
revno: 3254 [merge]
committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
branch nick: merge-5.5-security
timestamp: Mon 2011-12-12 12:28:23 +0100
message:
  merged 5.5->5.5-security
    ------------------------------------------------------------
    revno: 3097.116.62
    committer: kevin.lewis@oracle.com
    branch nick: mysql-5.5
    timestamp: Thu 2011-12-08 13:32:19 -0600
    message:
      Bug #13116225 LIVE DOWNGRADE CRASHES WITH INNODB_PAGE_SIZE=4K
      
      This bug ensures that a live downgrade from an InnoDB 5.6 with WL5756 and
      a database created with innodb-page-size=8k or 4k will make a version 5.5
      installation politely refuse to start. Instead of crashing or giving some
      indication about a corrupted database, it will indicate the page size
      difference.
      
      This patch takes only that part of the Wl5756 patch that is needed to
      protect against opening a tablespace that is stamped with a different
      page size.
      
      It also contains the change in dict_index_find_on_id_low() just in case
      a database with another page size is created by recompiling a pre-WL5756
      InnoDB.
    ------------------------------------------------------------
    revno: 3097.116.61
    committer: Jimmy Yang <jimmy.yang@oracle.com>
    branch nick: mysql-5.5
    timestamp: Wed 2011-12-07 18:44:52 -0800
    message:
      Fix Bug #13083023 - 60229: BROKEN COMPATIBILITY: ERROR WHILE CREATE TABLE
      WITH FOREIGN KEY CONSTRAI
      
      rb://844 approved by marko
    ------------------------------------------------------------
    revno: 3097.116.60
    committer: Inaam Rana <inaam.rana@oracle.com>
    branch nick: mysql-5.5
    timestamp: Wed 2011-12-07 09:12:53 -0500
    message:
      Bug#11759044 - 51325: DROPPING AN EMPTY INNODB TABLE TAKES A LONG TIME
      WITH LARGE BUFFER POOL
      
      (Note: this a backport of revno:3472 from mysql-trunk)
      
      rb://845
      approved by: Marko
      
        When dropping a table (with an .ibd file i.e.: with
        innodb_file_per_table set) we scan entire LRU to invalidate pages from
        that table. This can be painful in case of large buffer pools as we hold
        the buf_pool->mutex for the scan. Note that gravity of the problem does
        not depend on the size of the table. Even with an empty table but a
        large and filled up buffer pool we'll end up scanning a very long LRU
        list.
        
        The fix is to scan flush_list and just remove the blocks belonging to
        the table from the flush_list, marking them as non-dirty. The blocks
        are left in the LRU list for eventual eviction due to aging. The
        flush_list is typically much smaller than the LRU list but for cases
        where it is very long we have the solution of releasing the
        buf_pool->mutex after scanning 1K pages.
        
        buf_page_[set|unset]_sticky(): Use new IO-state BUF_IO_PIN to ensure
        that a block stays in the flush_list and LRU list when we release
        buf_pool->mutex. Previously we have been abusing BUF_IO_READ to achieve
        this.
    ------------------------------------------------------------
    revno: 3097.116.59
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5-bug13013970geom
    timestamp: Mon 2011-12-05 15:42:45 +0100
    message:
      Bug#13013970 MORE CRASHES IN FIELD_BLOB::GET_KEY_IMAGE
      
      The predicate is re-written from
      ((`test`.`g1`.`a` = geometryfromtext('')) or ...
      to
      ((`test`.`g1`.`a` = <cache>(geometryfromtext(''))) or ...
      
      The range optimizer calls save_in_field_no_warnings, in order to fetch keys.
      save_in_field_no_warnings returns 0 because of the cache wrapper,
      and get_mm_leaf() proceeded to call Field_blob::get_key_image()
      which accesses un-initialized data.
    ------------------------------------------------------------
    revno: 3097.116.58
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5-sighandler
    timestamp: Fri 2011-12-02 15:16:39 +0100
    message:
      Bug#11761576 post-push fix: HAVE_EXPLICIT_TEMPLATE_INSTANTIATION in header file broke Mac build
    ------------------------------------------------------------
    revno: 3097.116.57
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5-sighandler
    timestamp: Fri 2011-12-02 14:16:48 +0100
    message:
      Bug#11761576 54082: HANDLE_SEGFAULT MAKES USE OF UNSAFE FUNCTIONS
      
      handle_segfault is the signal handler code of mysqld.  however, it makes
      calls to potentially unsafe functions localtime_r, fprintf, fflush.
    ------------------------------------------------------------
    revno: 3097.116.56 [merge]
    committer: Inaam Rana <inaam.rana@oracle.com>
    branch nick: mysql-5.5
    timestamp: Thu 2011-12-01 09:47:12 -0500
    message:
      merge from mysql-5.1
        ------------------------------------------------------------
        revno: 2661.787.65
        committer: Inaam Rana <inaam.rana@oracle.com>
        branch nick: mysql-5.1
        timestamp: Thu 2011-12-01 09:41:52 -0500
        message:
          Bug#13414773 -INNODB_FAST_SHUTDOWN=2, ASSERT STATE == BUF_BLOCK_ZIP_PAGE
          
          We can have dirty pages during a fast shutdown. Relax the assertion.
    ------------------------------------------------------------
    revno: 3097.116.55
    committer: sayantan.dutta@oracle.com
    branch nick: 11746897
    timestamp: Thu 2011-12-01 18:54:29 +0530
    message:
      BUG #11746897 - 29508: PLEASE IMPLEMENT MYSQL-TEST-RUN.PL --STRACE-MASTER
      Includes fix for strace-client and restricted to strace and linux only.
      ******
    ------------------------------------------------------------
    revno: 3097.116.54
    committer: Luis Soares <luis.soares@oracle.com>
    branch nick: mysql-5.5
    timestamp: Thu 2011-12-01 00:54:54 +0000
    message:
      BUG#11745230
      
      Refactored the test case: hardened and extended it. Created test inc file
      to abstract the task of relocating binlogs.
      
      Also, disabled it on windows for not cluttering the test case any further,
      as it depends heavily on doing filesystem operations and path handling.
    ------------------------------------------------------------
    revno: 3097.116.53 [merge]
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5
    timestamp: Wed 2011-11-30 17:13:04 +0100
    message:
      NULL merge 5.1 => 5.5
        ------------------------------------------------------------
        revno: 2661.787.64
        committer: Tor Didriksen <tor.didriksen@oracle.com>
        branch nick: 5.1-sighandler
        timestamp: Wed 2011-11-30 17:11:13 +0100
        message:
          Bug#11761576 54082: HANDLE_SEGFAULT MAKES USE OF UNSAFE FUNCTIONS
          
          Post-push fix: build break on windows/optimized
    ------------------------------------------------------------
    revno: 3097.116.52 [merge]
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5
    timestamp: Wed 2011-11-30 15:56:38 +0100
    message:
      Mostly NULL-merge 5.1 => 5.5 of
      Bug#11761576 54082: HANDLE_SEGFAULT MAKES USE OF UNSAFE FUNCTIONS
        ------------------------------------------------------------
        revno: 2661.787.63
        committer: Tor Didriksen <tor.didriksen@oracle.com>
        branch nick: 5.1-sighandler
        timestamp: Wed 2011-11-30 15:39:29 +0100
        message:
          Bug#11761576 54082: HANDLE_SEGFAULT MAKES USE OF UNSAFE FUNCTIONS
          
          handle_segfault is the signal handler code of mysqld.  however, it makes
          calls to potentially unsafe functions localtime_r, fprintf, fflush.
    ------------------------------------------------------------
    revno: 3097.116.51 [merge]
    committer: Tor Didriksen <tor.didriksen@oracle.com>
    branch nick: 5.5
    timestamp: Wed 2011-11-30 10:45:23 +0100
    message:
      NULL merge 5.1 => 5.5
        ------------------------------------------------------------
        revno: 2661.787.62
        committer: Tor Didriksen <tor.didriksen@oracle.com>
        branch nick: 5.1
        timestamp: Tue 2011-11-29 15:52:47 +0100
        message:
          Build broken for gcc 4.5.1 in optimized mode.
          
          readline.cc: In function char* batch_readline(LINE_BUFFER*):
          readline.cc:60:9: error: out_length may be used uninitialized in this function
          log.cc: In function int find_uniq_filename(char*):
          log.cc:1857:8: error: number may be used uninitialized in this function
    ------------------------------------------------------------
    revno: 3097.116.50
    committer: Andrei Elkin <andrei.elkin@oracle.com>
    branch nick: mysql-5.5
    timestamp: Tue 2011-11-29 22:30:04 +0200
    message:
      reverting the initial patch for bug#13437900 for refinement.
    ------------------------------------------------------------
    revno: 3097.116.49
    committer: Andrei Elkin <andrei.elkin@oracle.com>
    branch nick: 5.5-bug13437900-repl_ignore_server_ids
    timestamp: Tue 2011-11-29 20:17:02 +0200
    message:
      Bug#13437900 - VALGRIND REPORTS A LEAK FOR REPL_IGNORE_SERVER_IDS
      
      There was memory leak when running some tests on PB2.
      The reason of the failure is an early return from change_master()
      that was supposed to deallocate a dyn-array.
      
      Fixed with relocating the dyn-array's destructor at ~LEX() that is
      the end of the session, per Gleb's patch idea.
      Two optimizations were done: the static buffer for the dyn-array to base on,
      and the array initialization is called precisely when it's necessary rather than
      per each CHANGE-MASTER as before.
       
    ------------------------------------------------------------
    revno: 3097.116.48 [merge]
    committer: Luis Soares <luis.soares@oracle.com>
    branch nick: mysql-5.5
    timestamp: Tue 2011-11-29 14:34:16 +0000
    message:
      BUG#11745230
      
      Automerged approved bzr bundle into latest mysql-5.5.
        ------------------------------------------------------------
        revno: 3097.120.1
        committer: Luis Soares <luis.soares@oracle.com>
        branch nick: mysql-5.5
        timestamp: Thu 2011-11-24 17:15:58 +0000
        message:
          BUG#11745230: 12133: MASTER.INDEX FILE KEEPS MYSQLD FROM STARTING IF
                        BIN LOG HAS BEEN MOVED
          
          When moving the binary/relay log files from one location to
          another and restarting the server with a different log-bin or
          relay-log paths, would cause the startup process to abort. The
          root cause was that the server would not be able to find the log
          files because it would consider old paths for entries in the
          index file instead of the new location.  What's even worse, the
          relative paths would not be considered relative to the path
          provided in log-bin and relay-log, but to mysql_data_dir.
                
          We fix the cases where the server contains relative paths. When
          the server is reading from the index file, it checks whether the
          entry contains relative paths. If it does, we replace it with the
          absolute path set in log-bin/relay-log option. Absolute paths
          remain unchanged and the index must be manually edited to
          consider the new log-bin and/or relay-log path (this should be
          documented). This is a fix for a GA version, that does not break
          behavior (that much).
                
          For development versions, we should go with Zhenxing's approach
          that removes paths altogether from index files.
    ------------------------------------------------------------
    revno: 3097.116.47 [merge]
    committer: Nirbhay Choubey <nirbhay.choubey@oracle.com>
    branch nick: mysql-5.5
    timestamp: Tue 2011-11-29 19:31:51 +0530
    message:
      Merge of patch for bug#11756764 from mysql-5.1.
        ------------------------------------------------------------
        revno: 2661.787.61
        committer: Nirbhay Choubey <nirbhay.choubey@oracle.com>
        branch nick: B11756764-5.1
        timestamp: Tue 2011-11-29 17:59:35 +0530
        message:
          Bug#11756764 48726: MYSQLD KEEPS CRASHING WITH SIGSEGV
                              WITH MYISAM_USE_MMAP ENABLED
          
          MySQL server can crash due to segmentation fault when
          started with myisam_use_mmap.
          
          The reason behind this being, while making a request to
          unmap (munmap) the previously mapped memory (mmap), the
          size passed was 7 bytes larger than the size requested at
          the time of mapping. This can eventually unmap the adjacent
          memory mapped block, belonging to some other memory-map pool.
          Hence the subsequent call to mmap can map a region which was
          still a valid memory mapped area.
          
          Fixed by removing the extra 7-byte margin which was erroneously
          added to the size, used for unmappping.
    ------------------------------------------------------------
    revno: 3097.116.46 [merge]
    committer: Luis Soares <luis.soares@oracle.com>
    branch nick: mysql-5.5-latest
    timestamp: Thu 2011-11-24 15:39:46 +0000
    message:
      BUG#13427949
      
      Automerged against latest mysql-5.5.
        ------------------------------------------------------------
        revno: 3097.119.1
        committer: Luis Soares <luis.soares@oracle.com>
        branch nick: mysql-5.5
        timestamp: Thu 2011-11-24 14:51:18 +0000
        message:
          BUG#13427949: CHANGE MASTER TO USER='' (EMPTY USER) CAUSES ERRORS ON VALGRING
            
          When passing an empty user to the connect function will cause
          valgrind warnings. Seems that the client code is not prepared
          to handle empty users. On 5.6 this can even be triggered by
          START SLAVE PASSWORD='...'; i.e., without setting USER='...' on
          the START SLAVE command (see WL#4143 for details on the new
          additional START SLAVE commands).
            
          To fix this, we disallow empty users when configuring the slave
          connection parameters (this decision might be revisited if the
          client code accepts empty users in the future).
    ------------------------------------------------------------
    revno: 3097.116.45 [merge]
    committer: Ashish Agarwal<ashish.y.agarwal@oracle.com>
    branch nick: mysql-5.5
    timestamp: Wed 2011-11-23 18:50:29 +0530
    message:
      BUG#11751793 - 42784: Merge from mysql-5.1 to mysql-5.5
        ------------------------------------------------------------
        revno: 2661.787.60
        committer: Ashish Agarwal<ashish.y.agarwal@oracle.com>
        branch nick: bug_93_5.1
        timestamp: Wed 2011-11-23 18:33:29 +0530
        message:
          BUG#11751793 - 42784: ARCHIVE TABLES CAUSE 100% CPU USAGE
                                AND HANG IN SHOW TABLE STATUS.
          
          ISSUE: Table corruption due to concurrent queries.
                 Different threads running insert and check
                 query leads to table corruption. Not properly locked,
                 rows are inserted in between check query.
          
          SOLUTION: In check query mutex lock is acquired
                    for a longer time to handle concurrent
                    insert and check query.
          
          NOTE: Additionally we backported the fix for CHECKSUM
                issue(bug#11758979).
    ------------------------------------------------------------
    revno: 3097.116.44
    committer: Build Team <MYSQL-RE_WW@oracle.com>
    branch nick: mysql-5.5
    timestamp: Wed 2011-11-23 12:33:59 +0100
    message:
      Bumped version  number to 5.5.20
    ------------------------------------------------------------
    revno: 3097.116.43
    tags: mysql-5.5.19, clone-5.5.19-build
    committer: Jimmy Yang <jimmy.yang@oracle.com>
    branch nick: mysql-5.5
    timestamp: Tue 2011-11-22 07:06:19 -0800
    message:
      Disable innodb_corrupt_bit.test on windows, issues with
      innodb_change_buffering_debug prevents creating the Dup Key scenario on
      windows