Download |Release Notes |Changelog |Overview of 5.3
Release date: 15 Feb 2012
For the highlights of this release, see the release notes.
The revision number links will take you to the revision's page on Launchpad. On Launchpad you can view more details of the revision and view diffs of the code modified in that revision.
Revision #3421 Tue 2012-02-14 16:52:56 +0200
Fix for causes 1093 error on Drupal
Problem was that now we can merge derived table (subquery in the FROM clause).
Fix: in case of detected conflict and presence of derived table "over" the table which cased the conflict - try materialization strategy.
Tue 2012-02-14 16:04:06 +0400
Review and push crashsafe GIS keys.
tests for RTree keys recovery.
[merge] Tue 2012-02-14 14:13:10 +0400
Merge fix for
Tue 2012-02-14 13:58:57 +0400
: Query containing IN subquery with OR in the where clause returns a wrong result
Mon 2012-02-13 23:46:57 -0800
Fixed .
If the flag 'optimize_join_buffer_size' is set to 'off' and the value of the system variable 'join_buffer_size' is greater than the value of the system variable 'join_buffer_space_limit' than no join cache can be employed to join tables of the executed query.
A bug in the function JOIN_CACHE::alloc_buffer allowed to use join buffer even in this case while another bug in the function revise_cache_usage could cause a crash of the server in this case if the chosen execution plan for the query contained outer join or semi-join operation.
Mon 2012-02-13 17:14:10 +0100
When we fail during slave thread initialisation, keep mi->run_lock locked until we have finished clean up.
Previously, the code released the lock without marking that the thread was running. This allowed a new slave thread to start while the old one was still in the middle of cleaning up, causing assertions and probably general mayhem.
[merge] Sun 2012-02-12 23:03:36 +0100
merge
Sun 2012-02-12 23:02:56 +0100
Use newly released HeidiSQL 7.0 in the MSI installer
[merge] Sat 2012-02-11 15:00:15 +0100
merge
[merge] Sat 2012-02-11 14:57:44 +0100
Fix for , - broken SSL connectivity for some connectors
[merge] Fri 2012-02-10 22:09:57 +0100
Merge fix for : Race condition in kill_threads_for_user().
[merge] Fri 2012-02-10 21:20:32 +0100
Merge fix for : Race condition in kill_threads_for_user().
[merge] Fri 2012-02-10 22:56:34 +0200
Automatic merge
Fri 2012-02-10 22:53:46 +0200
Fixed that prepared statements are properly igored (if possible) by query cache.
Thu 2012-02-09 23:35:26 +0200
Test case for
The bug itself is fixed by the patch for
Fri 2012-02-03 12:49:17 -0800
Made mrr related counters temporarily invisible until we agree upon their names.
Fri 2012-02-03 16:56:12 +0200
Fixed typos in Query Cache.
[merge] Fri 2012-02-03 13:32:29 +0200
Merge with 5.2
Fri 2012-02-03 12:24:55 +0200
Fixed DELETE issues of view over view over table.
Fri 2012-02-03 13:01:05 +0200
Moving to 5.3 + fixing INSERT of multi-table view.
Fri 2012-02-03 00:46:34 -0800
Back-ported the fix for bug #12831587 from mysql-5.6 code line.
The comment for the fix commit says:
Due to the changes required by ICP we first copy a row from the InnoDB format to the MySQL row buffer and then copy it to the pre-fetch queue. This was done for the non-ICP code path too. This change removes the double copy for the latter.
Thu 2012-02-02 22:22:27 -0800
Applied the patch from mysql-5.6 code line that fixed a significant performance drop for high concurrency bechmarks (bug #11765850
).
Here's the comment of the patch commit:
The bug is that the InnoDB pre-fetch cache was not being used in row_search_for_mysql(). Secondly the changeset that planted the bug also introduced some inefficient code. It would read an extra row, convert it to MySQL row format (for ICP==off), copy the row to the pre-fetch cache row buffer, then check for cache overflow and dequeue the row that was pushed if there was a possibility of a cache overflow.
[merge] Wed 2012-02-01 17:48:45 -0800
Merge.
[merge] Wed 2012-02-01 15:48:02 -0800
Merge 5.2->5.3 in preparation for the release of mariadb-5.3.4-rc.
Wed 2012-02-01 17:09:49 +0200
fix for
Problem was in try to check/use Item_direct_ref of derived view when we have to use real Item_field under it.
Tue 2012-01-31 21:12:26 +0100
sort status variables alphabetically in mysqld.cc
Mon 2012-01-30 20:34:47 +0400
: Loosescan reports different result than other semijoin methods
If LooseScan is used with quick select, require that quick select produces data in key order (this disables use of MRR, which can return data in arbitrary order).
[merge] Mon 2012-01-30 17:38:14 +0400
Merge fix for
Fri 2012-01-27 17:35:26 +0400
Sat 2012-01-28 01:12:45 -0800
Fixed
Applied the fix for bug #12546542 from the mysql-5.6 code line:
JOIN_CACHE::join_records forgot to reset JOIN_TAB::first_unmatched in some cases.
[merge] Fri 2012-01-27 19:23:08 -0800
Merge.
Fri 2012-01-27 19:01:26 -0800
Back-ported test cases for of mysql-5.6 code line. The bug could not be reproduced in the latest release of mariadb-5.3 as it was fixed by Sergey Petrunia when working on the problems concerning outer joins within in subqueries converted to semi-joins.
Fri 2012-01-27 17:51:40 +0400
Make testcase stable by adding --sorted_result for SHOW STATUS commands.
Thu 2012-01-26 14:20:34 +0400
Fix compile failure when built without query cache:
define QUERY_CACHE_DB_LENGTH_SIZE 0, just like it is done
with QUERY_CACHE_FLAGS_SIZE.
Thu 2012-01-26 12:22:02 +0400
Sort counters by name (will this make them show in the same order on all platforms?)
Wed 2012-01-25 22:05:20 +0400
: Wrong result (missing rows) with firstmatch+BNL, IN subquery, ...
Disable use of join cache when we're using FirstMatch strategy, and the join order is such that subquery's inner tables are interleaved with outer. Join buffering code is incapable of handling such join orders.
[merge] Wed 2012-01-25 18:36:57 +0400
Merge
Wed 2012-01-25 18:33:57 +0400
: Wrong result (extra rows) with loosescan and IN subquery
Mon 2012-01-23 15:14:13 +0200
Fixed creating limit for exists subquery.
Sun 2012-01-22 12:54:30 -0800
Back-ported test cases for , , , from subquery_sj of mysql-5.6 code line. The bugs could not be reproduced in the latest release of mariadb-5.3 as they were fixed either when the code of subquery optimization was back-ported from mysql-6.0 or later when some other bugs were fixed.
Sat 2012-01-21 20:58:23 -0800
Back-ported test cases for , , , from subquery_sj of mysql-5.6 code line. The bugs could not be reproduced in the latest release of mariadb-5.3 as they were fixed either when the code of subquery optimization was back-ported from mysql-6.0 or later when some other bugs were fixed.
Fri 2012-01-20 02:11:53 +0400
: Wrong result (missing rows) with join_cache_hashed+materialization+semijoin=on
equality substitution code was geared towards processing WHERE/ON clauses. that is, it assumed that it was doing substitions on the code that
wasn't attached to any particular join_tab yet
Thu 2012-01-19 23:44:43 +0400
: Wrong result (missing rows) with semijoin=on, firstmatch=on, ...
setup_semijoin_dups_elimination() would incorrectly set join_tab->do_firstmatch when the join order had outer tables interleaved with inner.
Thu 2012-01-19 13:46:59 +0100
Fix compiler warning on Windows.
[merge] Wed 2012-01-18 15:28:13 -0800
Merge
[merge] Wed 2012-01-18 14:19:28 -0800
Merge
Wed 2012-01-18 14:52:56 +0100
multi-delete should ignore semi-join internal temp tables, when looking for tables to delete from
[merge] Wed 2012-01-18 14:52:38 +0100
more problems found by PVS Studio
[merge] Mon 2011-11-14 00:21:22 +0100
5.2->5.3 merge
Wed 2012-01-18 12:53:50 +0200
Adjust test results after Monty's push of the new handler counter Handler_read_rnd_deleted.
Tue 2012-01-17 15:16:00 +0200
Removed unused code merged from 5.2. In 5.3 we fix this problem by checking if we put max/min group function without GROUP BY artificially in case of MODE_ONLY_FULL_GROUP_BY sql mode.
Fri 2012-01-13 14:35:49 +0200
Added Handler_read_rnd_deleted, number of deleted rows found with ha_read_rnd_first.
Wed 2012-01-11 10:35:16 +0200
Fix the comment.
Tue 2012-01-10 23:26:00 +0200
Fix for Wrong result with subquery in select list, EXISTS, constant MyISAM/Aria table.
Problem: When building the condition for JOIN::outer_ref_cond the optimizer forgot to take into account that this condition could depend on constant tables as well.
[merge] Tue 2012-01-10 18:20:56 +0400
Merge fix for
Sun 2012-01-08 14:43:14 +0400
: Crash in do_copy_not_null with semijoin=ON, firstmatch=ON, aggregate ...
Mon 2012-01-09 13:49:47 +0200
Fixed that --sorted-result in mysql-test-run also works for exec
Sun 2012-01-08 21:14:07 +0100
- possible deadlock in XtraDB async io subsystem on Windows.
Split IO threads into ones that handle only read completion and ones that handle only write completion, as it was originally done, but got lost with "completion port" patch. The reason we need to have dedicated read and dedicated write threads is that read completion routine can block waiting for write io to complete, and in rare cases where all io threads are handling async reads, it can deadlock.
Mon 2012-01-02 20:06:36 -0800
Fixed
The patch for erroneously removed a call of engine->set_thd() from Item_subselect::fix_fields().
Fri 2011-12-30 22:19:05 +0100
Make test results stable.
Fri 2011-12-30 20:22:52 +0100
Update version in configure.in (was forgotten in 5.3.3 release).
Fri 2011-12-30 11:34:29 +0100
Continuation of the efforts in previous cset.
Thu 2011-12-29 22:29:02 +0100
Make test results stable (they weren't, because filesort() used to read from a heap temptable, which uses pointers to records (that is, byte* pointers) as rowids.
This meant that for rows with the same sort key value, the order was determined by memory layout.
Wed 2011-12-28 12:12:48 +0400
Update test results.
[merge] Wed 2011-12-28 03:37:34 +0400
Merge
Tue 2011-12-20 00:55:32 +0400
: Incorrect result with outer join and full text match
Sun 2011-12-25 18:03:03 -0800
Changed a test case from join_cache.test to make it platform independent.
Sat 2011-12-24 08:55:10 -0800
Back-ported the patch of the mysql-5.6 code line that fixed several defects in the greedy optimization:
The greedy optimizer calculated the 'compare-cost' (CPU-cost) for iterating over the partial plan result at each level in the query plan as 'record_count / (double) TIME_FOR_COMPARE'
This cost was only used locally for 'best' calculation at each level, and not accumulated into the total cost for the query plan.
This fix added the 'CPU-cost' of processing 'current_record_count' records at each level to 'current_read_time' before it is used as 'accumulated cost' argument to recursive best_extension_by_limited_search() calls. This ensured that the cost of a huge join-fanout early in the QEP was correctly reflected in the cost of the final QEP.
To get identical cost for a 'best' optimized query and a straight_join with the same join order, the same change was also applied to optimize_straight_join() and get_partial_join_cost() 1. Furthermore to get equal cost for 'best' optimized query and a straight_join the new code substrcated the same '0.001' in optimize_straight_join() as it had been already done in best_extension_by_limited_search() 1. When best_extension_by_limited_search() aggregated the 'best' plan a plan was 'best' by the check :
'if ((search_depth == 1) || (current_read_time < join->best_read))'
The term '(search_depth == 1' incorrectly caused a new best plan to be collected whenever the specified 'search_depth' was reached - even if this partial query plan was more expensive than what we had already found.
Make equality propagation work correctly when done inside the OR branches
merge] Sat 2012-02-11 15:05:07 +0100
merge
Revision #2643.150.3 Sat 2012-02-11 03:25:49 +0100
fixes bug(s): Bug #930145
A recent change in the server protocol code broke SSL connection for some client libraries.
Protocol documentation (MySQL_Internals_ClientServer_Protocol) says that initial packet sent by client if client wants SSL, consists of client capability flags only (4 bytes or 2 bytes edependent on protocol versionl).
Some clients happen to send more in the initial SSL packet (C client, Python connector), while others (Java, .NET) follow the docs and send only client capability flags.
A change that broke Java client was a newly introduced check that frst client packet has 32 or more bytes. This is generally wrong, if client capability flags contains CLIENT_SSL.
Also, fixed the code such that read max client packet size and charset in the first packet prior to SSL handshake. With SSL, clients do not have to send this info, they can only send client flags.
This is now fixed such that max packet size and charset are not read prior to SSL handshake, in case of SSL they are read from the "complete" client authentication packet after SSL initialization.
Revision #3381.2.1 Fri 2012-02-10 21:19:12 +0100
Bug #910817: Race condition in kill_threads_for_user()
The code was accessing a pointer in a mem_root that might be freed by another concurrent thread. Fix by moving the access to be done while the LOCK_thd_data is held, preventing the memory from being freed too early.
Revision #2732.46.68 Fri 2012-02-03 10:31:30 +0200
Added 5.2 test result delimiter
Revision #2732.46.67 Fri 2012-02-03 10:28:23 +0200
Fix check of view updatability in case of underlying view changes its updatability.
For single table update/insert added deep check of single tables (single_table_updatable()). For multi-table view insert added additional check of target table (check_view_single_update).
Multi-update was correct.
Test suite for all cases added.
Revision #2732.46.66 Mon 2012-01-23 11:43:28 +0100
Revision #2732.46.65 Fri 2012-01-20 23:54:43 -0800
Fixed Bug #919427.
The function subselect_uniquesubquery_engine::copy_ref_key has to take into account that when EXPLAIN is processed the array of store_key object created for any TABLE_REF may contain elements for constant items. These items should be ignored by thefunction.
Revision #2732.46.64 Thu 2012-01-12 20:13:41 +0100
Bug #901693 dialog.c:perform_dialog treats every password prompt as first
Revision #2732.46.63 Thu 2012-01-12 20:13:22 +0100
Bug #893522 more problems found by PVS Studio
Revision #2732.46.62 Thu 2012-01-12 20:12:46 +0100
openpam compatibility
Revision #2732.46.61 Thu 2012-01-12 20:12:14 +0100
fixes for get_password():
on windows: don't hang when there's no console, that is, _getch() returns -1.
on windows: _getch() returns an int, not char. to distinguish between (char)255 and (int)-1
everywhere. isspace(pos[-1]) == ' ' never worked, isspace() returns a boolean, not a char. the never-worked loop was removed to preserve the existing behavior.
Revision #2732.46.60 Fri 2011-12-30 13:57:03 +0100
plugin renamed socket_peercred -> unix_socket.
test added.
Revision #2732.46.59 Thu 2011-12-29 22:52:13 +0100
on windows: don't link all plugins with mysqld, only do it for storage engines.
Revision #2732.46.58 Sat 2012-01-14 00:02:02 -0800
Back-ported the test case for bug #12616253 from mariadb-5.3 that was actually a duplicate of Bug #888456 fixed in mariadb-5.2.
Revision #2732.46.57 Fri 2012-01-13 19:00:50 -0800
Back-ported the fix and the test case for MySQL Bug #50257 from mariadb-5.3 code line.
Adjusted results for a few test cases.
Revision #2732.46.56 Fri 2012-01-13 12:23:19 -0800
Back-ported the test cases for bug #12763207 from mysql-5.6 code line into 5.2
Completed the fix for this bug.
Note: in 5.3 the affected 'if' statement in Item_in_subselect::single_value_transformer() starting with the condition (thd->variables.sql_mode & MODE_ONLY_FULL_GROUP_BY) should be removed altogether. The change from table.cc is not needed either.
This is because in 5.3
min/max transformation for subqueries are done at the optimization phase
evaluation of the expensive subqueries is done at the execution phase.
Added an EXPLAIN EXTENDED to the test case for bug #12329653.
Revision #2732.46.55 Tue 2012-01-10 19:26:47 +0100
MDEV-50 : Fix default compilation comment
Revision #2732.46.54 Tue 2012-01-10 19:23:00 +0100
Fix MDEV-49 : version_compile_machine server variable is 'unknown' for x64 builds
Revision #2732.46.53 Sun 2012-01-08 20:29:05 +0200
Fixed compiler and test failures found by buildbot
Revision #2732.46.52 [merge] Sat 2012-01-07 14:59:03 +0200
Merge
Revision #2732.51.1 Sat 2012-01-07 10:23:46 +0200
Fixed wrong merge
Revision #2732.46.51 Thu 2012-01-05 11:06:52 +0200
Fix of Bug #793589 Wrong result with double ORDER BY Problem was in caching 'eq_ref' dependency between calls of remove_const() for ORDER BY and GROUP BY lists.
Revision #2732.46.50 [merge] Wed 2012-01-04 17:22:06 +0200
Merge with 5.1
Revision #2643.143.68 Thu 2011-12-29 21:55:17 -0800
Fixed Bug #848652.
The cause of this bug was the same as for fixed for 5.3.
Wed 2011-12-21 13:23:15 +0200
Fixes "Server started with skip-aria crashes on an attempt to connect to it"
Revision #2732.46.49 Mon 2012-01-02 23:52:31 +0100
Fix embedded/windows tests- move COND_manager and LOCK_manager to sql_manager.cc, to prevent race condition that results into accessing already destroyed critical section
Revision #2732.46.48 Thu 2011-12-29 15:09:20 -0800
Fixed Bug #806057.
A table expression with a natural join or a USING clause is transformed into an equivalent expression with equi-join ON conditions. If a reference to a virtual column happened to occur only in these generated equi-join conditions then it was not erroneously marked in the TABLE::vcol_set bitmap. This could lead to wrong results for queries containing natural join expressions or USING clauses.
Revision #2732.46.47 Wed 2011-12-28 18:47:01 -0800
Fixed Bug #777654.
The method Item_sum_num::fix_fields() calculated the value of the flag Item_sum_num::maybe_null in some cases incorrectly.
Revision #2732.46.46 Tue 2011-12-27 19:13:53 -0800
Fixed Bug #879860.
The MIN/MAX optimization cannot be applied to a subquery if its WHERE clause contains a conjunctive condition depending on an outer reference.
Revision #2732.46.45 Tue 2011-12-27 13:19:13 -0800
Fixed Bug #904345.
The MIN/MAX optimizer code from the function opt_sum_query erroneously did not take into account conjunctive conditions that did not depend on any table, yet were not identified as constant items. These could be items containing rand() or PS/SP parameters. These items are supposed to be evaluated at the execution phase. That's why if such conditions can be extracted from the WHERE condition the MIN/MAX optimization is not applied as currently it is always done at the optimization phase.
(In 5.3 expensive subqueries are also evaluated only at the execution phase. So, if a constant condition with such subquery can be extracted from the WHERE clause the MIN/MAX optimization should not be applied in 5.3.)
IF an IN/ALL/SOME predicate with a constant left part is transformed into an EXISTS subquery the resulting subquery should not be considered uncacheable if the right part of the predicate is not uncacheable.
Backported the function dbug_print_item() from 5.3. The function is used only for debugging.
Revision #2732.46.44 [merge] Fri 2011-12-23 15:02:57 +0100
merge
Revision #2732.50.1 Thu 2011-12-22 15:50:33 +0100
Bug #906638 : Fixes to build oqgraph with boost 1.48
dijkstra_shortest_paths() needs a Graph as first parameter, in case of reverse_graph we now need to use its m_g member
use boost::tuples::tie() on all places where <<code.>tie()<</code>> was used . Reason -
fix the build with Visual Studio 10 SP1 (which includes std:tr1:tie, thus creating ambiguity)
Revision #2732.46.43 Thu 2011-12-22 11:07:04 +0100
compilation warning - unused variable
Revision #2732.46.42 Wed 2011-12-21 12:45:53 +0200
Supression condition made wider to cover some other system cases.
Revision #2732.46.41 Tue 2011-12-20 01:56:41 -0800
Fixed Bug #794005
The function st_table::mark_virtual_columns_for_write() did not take into account the fact that for any table the value of st_table::vfield is 0 when there are no virtual columns in the table definition.
Revision #2732.46.40 Mon 2011-12-19 14:55:30 -0800
Fixed Bug #906322
If the sorted table belongs to a dependent subquery then the function create_sort_index() should not clear TABLE:: select and TABLE::select for this table after the sort of the table has been performed, because these members are needed for the second execution of the subquery.
Fixed Item* Item_equal::get_first(JOIN_TAB *context, Item *field_item) to work correctly in the case where:
context!= NO_PARTICULAR_TAB, it points to a table within SJ-Materialization nest
field_item points to an item_equal that has a constant Item_field but does not have any fields from tables that are within semi-join nests.
The problem was that LooseScan execution code assumed that tab->key holds the index used for looseScan. This is only true when range or full index scan are used. In case of ref access, the index is in tab->ref.key (and tab->index==0 which explains how LooseScan passed tests with ref access: they used one index)
Fixed by setting/using loosescan_key, which always the correct index#.
Revision #3390.1.2 Wed 2012-01-25 18:27:34 +0400
Update handler status variables after the last commit.
Revision #3390.1.1 Mon 2012-01-23 23:35:52 +0400
Add MRR counters: Handler_mrr_init, Handler_mrr_extra_rowid_sorts, Handler_mrr_extra_key_sorts.
was going to be fed to make_join_select() which would take the condition apart and attach various parts of it to tables inside/outside semi-joins.
However, somebody added equality substition for ref access. That is, if we have a ref access on TBL.key=expr, they would do equality substition in 'expr'. This possibility wasn't accounted for.
Fixed equality substition code by adding a mode that does equality substition under assumption that the processed expression will be attached to a certain particular table TBL.
Revision #3381.1.1 Wed 2012-01-18 03:31:20 -0800
Fixed Bug #917990.
If the expression for a derived table of a query contained a LIMIT clause the estimate of the number of rows in this derived table returned by the EXPLAIN command could be badly off since the optimizer ignored the limit number from the LIMIT clause when getting the estimate.
The call of the method SELECT_LEX_UNIT->set_limit added in the code of mysql_derived_optimize() will be needed also in maria-5.5 where parameters in the LIMIT clause are supported.
Create/use do_copy_nullable_row_to_notnull() function for ref access, which is used when copying from not-NULL field in table that can be NULL-complemented to not-NULL field.
The problem was that const-table-reading code would try to evaluate MATCH() before init_ftfuncs() was called.
Fixed by making MATCH function "expensive" so that nobody tries to evaluate it at optimization phase.
This page is licensed: CC BY-SA / Gnu FDL