An error occurred fetching the project authors.
  1. 09 Sep, 2010 1 commit
    • Marc Alff's avatar
      Bug#56324 Race Condition while shutting down MySQL: "PSI_server" · 1e90c0a5
      Marc Alff authored
      Before this fix, the server could crash during shutdown,
      due to race conditions, that occured when killing the server.
      
      In particular, the performance schema instrumentation handle,
      PSI_server, and the performance schema itself would be cleaned up
      too soon, causing race conditions with a running kill server thread.
      
      The specifics of the race condition found are that:
      the main thread executing "PSI_server= NULL" can cause crashes in
      other threads still running, which are executing
      "if (PSI_server != NULL) PSI_server->xxx()"
      as part of the performance schema instrumentation.
      
      While the bug was reported for the kill server thread,
      in theory the same crash could happen with the signal thread,
      as found by code analysis.
      
      The correct fix would be to only shutdown the performance schema
      and set PSI_server to NULL after every other thread is guaranteed
      to be completed, including the kill_server_thread.
      
      However, due to the existing mysqld server design, this is not the case.
      See in particular bug number 56666.
      
      The work around used to fix this race condition is to simply not
      perform the call to shutdown_performance_schema() when the server exits,
      and to keep the PSI_server pointer unchanged.
      
      This will cause memory leaks to be reported by tools like valgrind,
      but no memory leak actually happen because the process is about to exit().
      
      As a result, the file mysql-test/valgrind.supp has been updated
      to filter out these false positive messages.
      
      This code has been tested with running in a loop the following
      tests in parallel, which have been known to fail with race conditions
      in the past:
      - rpl_change_master
      - binlog_max_extension
      - events_restart
      - rpl_heartbeat_basic
      and no crash of test failure has been seen with the changed code.
      1e90c0a5
  2. 06 Sep, 2010 1 commit
    • Mats Kindahl's avatar
      Bug #55966: "plugin" tests fail in 5.5 · 32206489
      Mats Kindahl authored
      On Solaris with version 3.4.6, the ha_example.so shared library is built
      with DTrace and the server is built without DTrace support. This occurs
      because dtrace.cmake disables DTrace support for 3.4.6, but still set
      HAVE_DTRACE, which causes probes_mysql.h to include probes_mysql_dtrace.h
      instead of probes_mysql_nodtrace.h.
      
      This patch fixes this by not setting HAVE_DTRACE on Solaris for GCC 3.4.6.
      32206489
  3. 02 Sep, 2010 2 commits
  4. 01 Sep, 2010 4 commits
  5. 31 Aug, 2010 3 commits
    • Alexander Nozdrin's avatar
      6fa0df0f
    • Alexander Nozdrin's avatar
      Bug#55980 Character sets: supplementary character _bin ordering is wrong · 8b4af2a2
      Alexander Nozdrin authored
      Problem:
      - ORDER BY for utf8mb4_bin, utf16_bin and utf32_bin returned
        results in a wrong order, because old functions
        (supporting only BMP range) were used to handle these collations.
      - Additionally, utf16_bin did not sort supplementary characters
        between U+D700 and U+E000, as WL#1213 specification specified.
      
      include/m_ctype.h:
        Adding prototypes.
      mysql-test/include/ctype_filesort2.inc:
        Adding a new shared test file.
      mysql-test/t/ctype_utf8mb4.test:
        Adding tests.
      strings/ctype-ucs2.c:
        - Fixing my_strncoll[sp]_utf16_bin to compare
          binary representation instead of code points,
          to make columns with indexes sort correct.
        - Fixing my_collation_handler_utf32_bin and
          my_collation_handler_utf16_bin to use new
          functions.
      strings/ctype-utf8.c:
        - Adding my_strnxfrm[len]_unicode_fill_bin()
          to handle utf8mb4_bin, utf16_bin and utf32_bin,
          using 3 bytes per weight.
          This function also performs special reordering in case of utf16_bin.
        - Fixing my_collation_utf8mb4_bin handler to use the
          new function.
      8b4af2a2
    • Alexander Nozdrin's avatar
      Auto-merge from mysql-5.5-merge. · e0622c5a
      Alexander Nozdrin authored
      e0622c5a
  6. 30 Aug, 2010 10 commits
    • Gleb Shchepa's avatar
      af604962
    • Gleb Shchepa's avatar
      Bug #53034: Multiple-table DELETE statements not accepting · 07024ab7
      Gleb Shchepa authored
                  "Access compatibility" syntax
      
      The "wild" "DELETE FROM table_name.* ... USING ..." syntax
      for multi-table DELETE statements is documented but it was
      lost in the fix for the bug 30234.
      
      The table_ident_opt_wild parser rule has been added
      to restore the lost syntax.
      
      
      mysql-test/r/delete.result:
        Test case for bug #53034.
      mysql-test/t/delete.test:
        Test case for bug #53034.
      sql/sql_yacc.yy:
        Bug #53034: Multiple-table DELETE statements not accepting
                    "Access compatibility" syntax
        
        The table_ident_opt_wild parser rule has been added
        to restore the lost syntax.
        Note: simple extending of table_ident with opt_wild in
        the table_alias_ref rule is not acceptable, because
        a) it adds one conflict more and b) this conflict resolves
        in the inappropriate way.
      07024ab7
    • Alexander Nozdrin's avatar
      Auto-merge from mysql-5.5-merge. · 8a50a7a2
      Alexander Nozdrin authored
      8a50a7a2
    • Magnus Blåudd's avatar
      Bug#56117 ha_ndbcluster_tables.h included by mysqldump but not used · 90eef290
      Magnus Blåudd authored
       - Remove include of ha_ndbcluster_tables.h from mysqldump.c
      90eef290
    • Dmitry Shulga's avatar
      Follow-up for Bug#29751: FLUSH LOGS doesn't create -old file, · 9a9e7a2f
      Dmitry Shulga authored
      so test case has to be updated.
      9a9e7a2f
    • Alexander Nozdrin's avatar
      Auto-merge from mysql-5.5. · e2dd2e73
      Alexander Nozdrin authored
      e2dd2e73
    • Alexander Nozdrin's avatar
      c5ca4142
    • Alexander Nozdrin's avatar
      Auto-merge from mysql-5.5-bugfixing. · bef62cb6
      Alexander Nozdrin authored
      bef62cb6
    • Alexey Kopytov's avatar
      Automerge. · c2a3d3c5
      Alexey Kopytov authored
      c2a3d3c5
    • unknown's avatar
      Bug #54579 Wrong unsafe warning for INSERT DELAYED in SBR · 3b1a7ebb
      unknown authored
            
      The lock_type is upgrade to TL_WRITE from TL_WRITE_DELAYED for
      INSERT DELAYED when inserting multi values in one statement.
      It's safe. But it causes an unsafe warning in SBR.
            
      Make INSERT DELAYED safe by logging it as INSERT without DELAYED.
      
      
      mysql-test/extra/binlog_tests/binlog_insert_delayed.test:
        Updated the test file to test multi INSERT DELAYED statement
        is no longer causes an unsafe warning and binlogged as INSERT
        without DELAYED.
      mysql-test/extra/rpl_tests/create_recursive_construct.inc:
        Updated for the patch of bug#54579.
      mysql-test/suite/binlog/r/binlog_row_binlog.result:
        Updated for the patch of bug#54579.
      mysql-test/suite/binlog/r/binlog_statement_insert_delayed.result:
        Test result for BUG#54579.
      mysql-test/suite/binlog/r/binlog_stm_binlog.result:
        Updated for the patch of bug#54579.
      mysql-test/suite/binlog/r/binlog_unsafe.result:
        Updated for the patch of bug#54579.
      mysql-test/suite/binlog/t/binlog_unsafe.test:
        Updated for the patch of bug#54579.
      sql/sql_insert.cc:
        Added code to genetate a new query string for removing
        DELAYED keyword for multi INSERT DEALAYED statement.
      sql/sql_yacc.yy:
        Added code to record the DELAYED keyword position and remove the setting
        of unsafe statement for INSERT DELAYED statement
      3b1a7ebb
  7. 28 Aug, 2010 1 commit
  8. 27 Aug, 2010 6 commits
    • Marc Alff's avatar
      local merge · 3b56ef04
      Marc Alff authored
      3b56ef04
    • Sergey Vojtovich's avatar
      Merge 5.1-bugteam to 5.5-merge. · 98df5f7c
      Sergey Vojtovich authored
      98df5f7c
    • Alexey Kopytov's avatar
      Bug #54465: assert: field_types == 0 || field_types[field_pos] · 36e234a8
      Alexey Kopytov authored
                  == MYSQL_TYPE_LONGLONG
      
      A MIN/MAX() function with a subquery as its argument could lead
      to a debug assertion on debug builds or wrong data on release
      ones.
      
      The problem was a combination of the following factors:
      
      - Item_sum_hybrid::fix_fields() might use the argument
      (args[0]) to calculate 'hybrid_field_type' which was later used
      to decide how the data should be sent to the client.
      
      - Item_sum::make_field() might use the argument again to
      calculate the field's type when sending result set metadata to
      the client.
      
      - The argument could be changed in between these two calls via
        Item::set_arg() leading to inconsistent metadata being
        reported.
      
      Here is what was happening for the bug's test case:
      
      1. Item_sum_hybrid::fix_fields() calculates hybrid_field_type
      as MYSQL_TYPE_LONGLONG based on args[0] which is an
      Item::SUBSELECT_ITEM at that time.
      
      2. A temporary table is created to execute the
      query. create_tmp_field_from_item() creates a Field_long object
      according to the subselect's max_length.
      
      3. The subselect item in Item_sum_hybrid is replaced by the
      Item_field object referencing the newly created Field_long.
      
      4. Item_sum::make_field() rightfully returns the
      MYSQL_TYPE_LONG type when calculating the result set metadata.
      
      5. When sending the actual data, Item::send() relies on the
      virtual field_type() function which in our case returns
      previously calculated hybrid_field_type == MYSQL_TYPE_LONGLONG.
      
      It looks like the only solution is to never refer to the
      argument's metadata after the result metadata has been
      calculated in fix_fields(), since the argument itself may be
      different by then. In this sense, Item_sum::make_field() should
      never be used, because it may rely on the argument's metadata
      and is only called after fix_fields(). The "default"
      implementation in Item::make_field() should be used instead as
      it relies only on field_type(), but not on the argument's type.
      
      Fixed by removing Item_sum::make_field() so that the superclass
      implementation Item::make_field() is always used.
      
      mysql-test/r/func_group.result:
        Added a test case for bug #54465.
      mysql-test/t/func_group.test:
        Added a test case for bug #54465.
      sql/item_sum.cc:
        Removed Item_sum::make_field() so that the superclass
        implementation Item::make_field() is always used.
      sql/item_sum.h:
        Removed Item_sum::make_field() so that the superclass
        implementation Item::make_field() is always used.
      36e234a8
    • Ramil Kalimullin's avatar
      Fix for bug #54253: memory leak when using I_S plugins w/o deinit method · e433cb33
      Ramil Kalimullin authored
      Free memory allocated by the server for all plugins,
      with or without deinit() method.
      
      e433cb33
    • Sergey Vojtovich's avatar
      e1af6730
    • Sergey Vojtovich's avatar
      BUG#52821 - plugin_ftparser.h and plugin_audit.h are · 8f773e6f
      Sergey Vojtovich authored
                  not tested by ABI check
      
      plugin_audit.h and plugin_ftparser.h are now subject
      for ABI check. plugin.h is now tested implicitly.
      
      Also fixed broken ABI check cmake rules.
      
      Makefile.am:
        plugin_audit.h and plugin_ftparser.h are now subject
        for ABI check. plugin.h is now tested implicitly.
      cmake/abi_check.cmake:
        plugin_audit.h and plugin_ftparser.h are now subject
        for ABI check. plugin.h is now tested implicitly.
        
        Also fixed broken ABI check rules: -DMYSQL_ABI_CHECK is
        compiler (not cmake) definition, incorrect definitions
        were passed to do_abi_check.cmake for abi_check_all rule.
      cmake/do_abi_check.cmake:
        Inform sources that we do ABI check.
      include/mysql/plugin.h.pp:
        plugin.h is now tested implicitly.
      include/mysql/plugin_audit.h.pp:
        plugin_audit.h is now subject for ABI check.
      include/mysql/plugin_ftparser.h.pp:
        plugin_ftparser.h is now subject for ABI check.
      8f773e6f
  9. 26 Aug, 2010 12 commits
    • Marc Alff's avatar
      local merge · c723902d
      Marc Alff authored
      c723902d
    • Christopher Powers's avatar
      Bug# 53874 "SETUP_INSTRUMENTS.TIMED='NO' should not change TIMER_WAIT · c696ec92
      Christopher Powers authored
      Handle combined instrument states of ENABLED and/or TIMED:
      
      ENABLED TIMED
      1 1 Aggregate stats, increment counter
      1 0 Increment counter
      0 1 Do nothing
      0 0 Do nothing 
      
      storage/perfschema/pfs.cc:
        Aggregate stats only if state is both ENABLED and TIMED. If ENABLED
        but not TIMED, only increment the value counter.
      storage/perfschema/pfs_stat.h:
        Split aggregate and counter increment into separate
        methods for performance.
      c696ec92
    • Alexander Nozdrin's avatar
    • Marc Alff's avatar
      local merge · d0dea5e4
      Marc Alff authored
      d0dea5e4
    • Marc Alff's avatar
      Implemented code review comments · e965de76
      Marc Alff authored
      e965de76
    • Bjorn Munch's avatar
      44d221d7
    • Alexey Kopytov's avatar
      Automerge. · 1e854f9c
      Alexey Kopytov authored
      1e854f9c
    • Alexander Barkov's avatar
      Bug#42511 mysqld: ctype-ucs2.c:2044: my_strnncollsp_utf32: Assertion (tlen % 4) == 0' fai · 7afb648f
      Alexander Barkov authored
      Problem: trailing spaces were stripped using 8-bit code,
      so the truncation result length was incorrect, which led
      to an assertion failure.
      Fix: using multi-byte safe code.
      7afb648f
    • Alexey Kopytov's avatar
      Automerge. · b1824004
      Alexey Kopytov authored
      b1824004
    • Sergey Vojtovich's avatar
      d61b21c3
    • Evgeny Potemkin's avatar
      Bug #55656: mysqldump can be slower after bug 39653 fix. · 94100243
      Evgeny Potemkin authored
      After fix for bug 39653 the shortest available secondary index was used for
      full table scan. Primary clustered key was used only if no secondary index
      can be used. However, when chosen secondary index includes all fields of the
      table being scanned it's better to use primary index since the amount of
      data to scan is the same but the primary index is clustered.
      Now the find_shortest_key function takes this into account.
      
      
      mysql-test/suite/innodb/r/innodb_mysql.result:
        Added a test case for the bug#55656.
      mysql-test/suite/innodb/t/innodb_mysql.test:
        Added a test case for the bug#55656.
      sql/sql_select.cc:
        Bug #55656: mysqldump can be slower after bug #39653 fix.
        The find_shortest_key function now prefers clustered primary key
        if found secondary key includes all fields of the table.
      94100243
    • Marc Alff's avatar
      Bug#55873 short startup options do not work in 5.5 · 6f0e7b9d
      Marc Alff authored
      Before this fix, the server did not recognize 'short' (as in -a)
      options but only 'long' (as in --ansi) options
      in the startup command line, due to earlier changes in 5.5
      introduced for the performance schema.
      
      The root cause is that handle_options() did not honor the
      my_getopt_skip_unknown flag when parsing 'short' options.
      
      The fix changes handle_options(), so that my_getopt_skip_unknown is
      honored in all cases.
      
      Note that there are limitations to this,
      see the added doxygen documentation in handle_options().
      
      The current usage of handle_options() by the server to
      parse early performance schema options fits within the limitations.
      This has been enforced by an assert for PARSE_EARLY options, for safety.
      6f0e7b9d