The most recent release in the MariaDB 5.5 series is:MariaDB 5.5.68 Download Now
Download |Release Notes |Changelog |[Overview of 5.5
Release date: 16 Mar 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 #3327 Wed 2012-03-14 19:47:15 +0100
another fix for --innodb-trx* name conflict
Tue 2012-03-13 14:43:34 +0100
feedback plugin is linked dynamically in bintar
Tue 2012-03-13 13:34:24 +0100
alternative method of resolving addresses for safemalloc and crash handler. don't link with libbfd, exec addr2line, if it's available at run time
Tue 2012-03-13 13:31:21 +0100
fix SphinxSE version as reported in I_S.PLUGINS
Tue 2012-03-13 13:29:44 +0100
monty's cleanup of my_thr_init.c and collateral changes
Tue 2012-03-13 13:28:08 +0100
broad suppression for dlsym "memory leak" - same as for dlclose
Tue 2012-03-13 13:27:14 +0100
disable EXTRA_DEBUG in non-dbug builds
[merge] Mon 2012-03-12 00:45:18 +0200
Merged the implementation of LIMIT ROWS EXAMINED into .
Sun 2012-03-11 14:39:20 +0200
Implementation of LIMIT ROWS EXAMINED
Fri 2012-03-09 09:20:45 +0100
fix uninitialized warning in mysql-test-run.pl make the test to require sphinx 2.0.4 or later
[merge] Fri 2012-03-09 08:06:59 +0100
merge with mysql-5.5.21
Thu 2012-03-08 13:06:28 +0100
remove .bzr-mysql directory - it's only used by proprietory bzr-mysql plugin
Wed 2012-03-07 17:38:47 +0200
Upgraded sphinx to version 2.0.4
Fixed memory leaks and compiler warnings in ha_sphinx.cc
Added HA_MUST_USE_TABLE_CONDITION_PUSHDOWN so that an engine can force index condition to be used
[merge] Tue 2012-03-06 20:46:07 +0100
5.3 merge
[merge] Mon 2012-03-05 22:33:46 -0800
Merge.
Mon 2012-03-05 21:55:25 +0100
compilation warning: unused variable
Mon 2012-03-05 21:48:06 +0100
: INSTALL PLUGIN SONAME
Mon 2012-03-05 17:44:26 +0100
updates after writing
Sat 2012-03-03 13:22:49 -0800
Supported extended keys () for innodb_plugin.
[merge] Sat 2012-03-03 09:16:30 +0100
Merge
[merge] Fri 2012-03-02 22:52:03 -0800
Merge
Fri 2012-03-02 08:32:16 +0100
misc test/result fixes
[merge] Fri 2012-03-02 07:45:06 +0100
Merge : @@skip_replication, into latest for push
Thu 2012-03-01 16:06:27 +0100
replicate_events_marked_for_skip does not exist in embedded.
[merge] Thu 2012-03-01 14:22:22 -0800
Merge 5.3->5.5.
Wed 2012-02-29 23:28:16 -0800
Fixed .
Thu 2012-03-01 17:04:57 +0100
misc test/result fixes
Wed 2012-02-29 21:55:53 +0100
Make Tc_log_page_size status variable use SHOW_LONG_NOFLUSH.
Otherwise XA crashes after FLUSH STATUS as log page size suddenly becomes 0.
Wed 2012-02-29 21:55:33 +0100
fixing pam plugin to compile again
Wed 2012-02-29 21:55:04 +0100
pbxt suite is now a main-pbxt overlay
Thu 2012-02-23 09:24:11 +0100
instead of having win/notwin tests that only differ in results, use one test with two combinations (win/unix), where only one is enabled.
Apply this technique to mysqld--help.
Thu 2012-02-23 09:18:48 +0100
don't even try to run xtradb-only tests with innodb, use have_xtradb.combinations.
Thu 2012-02-23 07:52:27 +0100
remove few .require files and one duplicate have_*inc file.
move variable tests from main to sys-vars
Thu 2012-02-23 07:50:43 +0100
HAVE_STRNDUP check for pam plugin
Thu 2012-02-23 07:50:11 +0100
overlay support for mysql-test-run and mysqltest
mysql-test-run auto-disables all optional plugins.
Tue 2012-02-07 18:53:33 +0100
making more use of My::Suite object
Tue 2012-02-07 17:18:41 +0100
small cleanup
Tue 2012-02-07 16:22:36 +0100
allow suite.pm to skip combinations that originate from test/include files.
Mon 2012-02-06 23:16:21 +0100
mtr: support for rdiff files
Mon 2012-02-06 22:55:17 +0100
per-combination result files
Mon 2012-02-06 21:36:56 +0100
per-file combinations
Mon 2012-02-06 20:29:21 +0100
cleanup
Mon 2012-02-06 20:29:13 +0100
make %suites hash local to mtr_cases.pm
Mon 2012-02-06 20:28:56 +0100
move --secure-file-priv from hardcoded to a template. remove redundant suite.opt
Mon 2012-02-06 18:42:18 +0100
remove few obscure, unused, or misused mtr features
Mon 2012-02-06 16:29:53 +0100
remove few hard-coded checks from mtr
Mon 2012-02-06 16:26:12 +0100
added plugin/auth_pam/CMakeLists.txt
Thu 2012-03-01 16:24:59 +0100
fixes
Assertion: prebuilt->sql_stat_start || trx->conc_state == 1 failed at row0sel.c:3933
DELETE IGNORE should not ignore deadlocks
Wed 2012-02-29 18:25:25 +0100
fixes
Fix build error on Ubuntu 11.10, if systemtap is installed.
The error is due to conflict between min/max macros in my_global.h and system header < limits>, indirectly included via generated probes_mysql_dtrace.h
Temporarily undefined min/max for the inclusion of the probes_mysq_dtrace.h
Tue 2012-02-28 18:53:05 +0100
Update copyright notices
[merge] Tue 2012-02-28 13:16:17 +0100
merge threadpool
Mon 2012-02-27 19:54:30 +0100
precache some more system checks on Windows
[merge] Sun 2012-02-26 16:11:44 +0100
Merge : non-blocking client library into MariaDB.
Thu 2012-02-23 15:42:21 +0100
after-merge fixes.
Sat 2012-02-25 16:13:24 +0100
make the test result independent from sizeof(void*)
This task implements a new clause LIMIT ROWS EXAMINED as an extention to the ANSI LIMIT clause. This extension allows to limit the number of rows and/or keys a query would access (read and/or write) during query execution.
Revision #2502.541.1 Mon 2012-03-05 20:32:28 -0800
Fixed Bug #946055.
The function create_hj_key_for_table() that builds the descriptor of the hash join key to access a table of a materialized subquery must ignore any equi-join predicate depending on the tables not belonging to the subquery.
Revision #2502.513.73 [merge] Mon 2012-03-05 22:00:24 +0200
Automatic merge
Revision #2502.540.2 Mon 2012-03-05 21:59:00 +0200
Ensure that we mark all processed tables as 'properly closed'.
This is needed as last log entry may be a DDL that is not processed and then a table may be left in 'not properly closed state' even if information is correct in it.
Tue 2012-02-28 23:18:52 +0200
Fixed "Querying myisam table metadata while 'alter table..enable keys' is running may corrupt the table"
Fixed wrong mutex order bug in Aria when flush_log_for_bitmap() was called when table is not yet marked for change.
Revision #2502.513.72 Mon 2012-03-05 15:48:12 +0200
Fix for Bug #944504
Problem is that subquery execution can't be called during prepare/optimize phase.
Also small fix for subquery test suite.
Revision #2502.513.71 Sun 2012-03-04 22:38:17 -0800
Fixed Bug #944782.
This bug in the function JOIN::drop_unused_derived_keys() could leave the internal structures for a materialized derived table in an inconsistent state. This led to a not quite correct EXPLAIN output when no additional key had been created to access the table. It also may lead to more serious consequences: so, the test case added with this fix caused a crash in mariadb-5.5.20.
Revision #3307.1.1 [merge] Fri 2012-03-02 15:03:20 -0800
Merge MWL#247 from mariadb 5.3 -> mariadb 5.5.
Revision #2502.539.6 Tue 2012-02-28 13:03:10 -0800
Addressed all review feedbacks for MWL#247.
[merge] Mon 2012-02-27 21:23:12 -0800
Merge.
[merge] Wed 2012-02-22 13:04:58 -0800
Merge.
Sun 2012-01-29 14:35:30 -0800
Fixed .
When working on I forgot to adjust the function create_hj_key_for_table() that created a key definition for hash join keys. The modified function must set the values of the fields ext_key_parts, ext_key_flags, ext_key_part_map added to the key definition structure in .
Tue 2012-01-24 23:34:02 -0800
Fixed LP bug #921167.
The fields ext_key_flags and ext_key_part_map must be initialized for any key, even for a MyISAM key that never is considered by the optimizer as one extended by hidden components.
[merge] Tue 2012-01-24 21:12:02 -0800
Merge.
Wed 2012-01-18 19:38:03 -0800
Added a test case for .
Revision #3283.1.2 Thu 2012-03-01 13:12:49 +0100
Small cleanups:
use thd->in_active_multi_stmt_transaction() for @@in_transaction, not THD flag directly
use common error_if_in_trans_or_substatement() function for all 4 binlog variables that cannot be changed inside statement or transaction.
Revision #3283.1.1 [merge] Thu 2012-03-01 12:41:49 +0100
Merge MWL#234: @@skip_replication feature to MariaDB 5.5.
Revision #2502.536.4 Tue 2011-08-16 11:51:02 +0200
MWL#234: Implement option to switch between master-side and client-side filtering of @@skip_replication events.
Mon 2011-08-15 10:05:14 +0200
: Add MTR tests for SESSION/GLOBAL semantics of new system variables.
Fri 2011-08-12 13:18:34 +0200
: After-review fixes, including better names for the new system variables.
Thu 2011-08-11 11:38:52 +0200
: Support for marking binlog events to not be replicated, and for telling slaves not to replicate events with such mark
This bug appeared after the patch for bug 939009 that in the function merge_key_fields forgot to reset a proper value for the val field in the result of the merge operation of the key field created for a regular key access and the key field created to look for a NULL key.
Adjusted the results of the test case for bug 939009 that actually were incorrect.
Revision #2502.513.69 Tue 2012-02-28 15:41:55 +0100
Bug #938977 - Query performance with join/index super slow on MariaDB 5.3.4 RC
make sure that stored routines are evaluated (that is, de facto - cached) in convert_const_to_int().
revert the fix for Bug #806943 because it cannot be repeated anymore.
add few tests for convert_const_to_int()
Revision #2502.513.68 [merge] Tue 2012-02-28 15:04:31 +0100
merge
Revision #2502.528.10 [merge] Tue 2012-02-28 13:50:30 +0200
Automatic merge
Tue 2012-02-28 13:39:02 +0200
Fixed "Assertion `page->size <= share->max_index_block_size'"
The issue was that Aria allowed too long keys to be created (so that the internal buffer was not big enough to hold the whole key).
Key lengths is now limited to HA_MAX_KEY_LENGTH (1000), as for MyISAM.
Revision #2502.513.67 [merge] Sun 2012-02-26 03:13:33 -0800
Merge.
Revision #2502.534.2 [merge] Sun 2012-02-26 02:42:45 -0800
Merge 5.2->5.3
Sat 2012-02-25 17:10:07 -0800
Fixed .
The field key_cache_mem_size of the KEY_CACHE structure must be initialized in the function init_key_cache() and updated in the function resize_key_cache().
Sat 2012-02-25 09:03:06 +0200
Fix of (also and )
Cause of the bug is uninitialized items before evaluation HAVING clasue in case of empty result.
Sun 2012-02-26 00:19:07 -0800
Rolled back the patch for bug 791761.
A better fix for this bug will be pulled from mariadb-5.2.
Revision #2502.513.66 Sun 2012-02-26 11:44:52 +0400
Bump the version number.
Revision #2502.513.65 [merge] Fri 2012-02-24 18:35:58 -0800
Merge.
Revision #2502.533.1 Fri 2012-02-24 16:50:22 -0800
Fixed Bug #939009.
The result of materialization of the right part of an IN subquery predicate is placed into a temporary table. Each row of the materialized table is distinct. A unique key over all fields of the temporary table is defined and created. It allows to perform key look-ups into the table. The table created for a materialized subquery can be accessed by key as any other table. The function best_access-path search for the best access to join a table to a given partial join. With some where conditions this function considers a possibility of a ref_or_null access. If such access employs the unique key on the temporary table then when estimating the cost this access the function tries to use the array rec_per_key. Yet, such array is not built for this unique key. This causes a crash of the server.
Rows returned by the subquery that contain nulls don't have to be placed into temporary table, as they cannot be match any row produced by the left part of the subquery predicate. So all fields of the temporary table can be defined as non-nullable. In this case any ref_or_null access to the temporary table does not make any sense and it does not make sense to estimate such an access.
The fix makes sure that the temporary table for a materialized IN subquery is defined with columns that are all non-nullable. The also ensures that any row with nulls returned by the subquery is not placed into the temporary table.
Revision #2502.513.64 Sat 2012-02-25 01:42:28 +0400
Update test results.
Revision #2502.513.63 Fri 2012-02-24 22:42:37 +0400
Bug #938131: Subquery materialization is not used in CREATE TABLE SELECT
Enable subquery materialization for CREATE TABLE ... SELECT.
Revision #2502.513.62 [merge] Fri 2012-02-24 20:07:12 +0400
Merge 5.2->5.3
Revision #2502.528.7 [merge] Fri 2012-02-24 17:21:44 +0200
Automatic merge
Fri 2012-02-24 17:01:47 +0200
Fix for : MariaDB crashes on a select with long varchar and blob fields
Problem was a crash in internal temporary (Maria) files when row length exceeded 65535
Wed 2012-02-22 00:10:39 -0800
Back-ported the fix and test cases for bugs #59487 and #43368 from the mysql-5.6 code line.
Revision #2502.513.61 [merge] Fri 2012-02-24 17:13:04 +0400
Merge fix for Bug #934597
Revision #2502.532.1 Fri 2012-02-24 17:09:13 +0400
Bug #934597: Assertion `! is_set()' failed in Diagnostics_area::set_ok_status(THD...
After the exec_const_cond->val_int() call, check for error and return.
(if we don't do it, we will eventually hit an error when trying to set status OK in the diagnostics area, which already has an error status).
Revision #3168.1.60 Mon 2012-02-27 19:53:49 +0100
Remove libevent from sources - not needed now
Revision #3168.1.59 [merge] Mon 2012-02-27 19:32:44 +0100
merge 5.5
Revision #3281.1.1 Mon 2012-02-27 19:20:18 +0100
because of the high cost associated with fake symdir resolution, disable symbolic-links on Windows by default. Real symlinks (Vista+) as well as NTFS junctions (prior to Vista) do not require this parameter
Revision #3168.1.58 Fri 2012-02-17 23:33:18 +0100
Simplify thd_wait_begin. given how seldom they are called, calling current_thd one more time is not going to be anything performance relevant.
Also use thd_wait_begin/end for thr_lock and sync callbacks.
Revision #3168.1.57 Fri 2012-02-17 23:27:15 +0100
Added copiright, some more comments
Revision #3168.1.56 Fri 2012-02-17 23:23:54 +0100
Store callback instance in the connection structure, to call CallbackMayRunLong on long waits (currently binlog only)
Also add copyright notice.
Revision #3168.1.55 Fri 2012-02-17 03:34:33 +0100
fix windows embedded (default thread handling ==pool-of-threads does not work in embedded)
Revision #3168.1.54 Thu 2012-02-16 21:07:22 +0100
use poof-of-threads as default for thread_handling on Windows
Revision #3168.1.53 Thu 2012-02-16 18:12:40 +0100
Only synchronous disk reads should use thd_wait_begin with THD_WAIT_DISKIO
Revision #3168.1.52 [merge] Thu 2012-02-16 17:33:37 +0100
merge from 5.5
Revision #3168.1.51 Thu 2012-02-16 16:59:04 +0100
address second round review comments
merge] Wed 2012-02-08 11:18:55 +0100
merge
Revision #3168.1.49 Sat 2012-01-28 01:09:28 +0100
some more whitespace, remove pending_thread_start_count. increment counters (thread_group->count, thread_group->active_thread_count) whenever mysql_create_thread returns success.
Revision #3168.1.48 Fri 2012-01-27 21:24:17 +0100
Fix test case - result file needs one-thread-per-connection
Revision #3168.1.47 Fri 2012-01-27 19:52:53 +0100
Threadpool : Rest of monty's review
merge] Fri 2012-01-27 00:40:12 +0100
merge
Revision #3168.1.45 Fri 2012-01-27 00:39:23 +0100
close callbacks prior to closing connection to avoid potential race when e.g timer callback and connection_destroy run in parallel
merge] Thu 2012-01-26 20:09:25 +0100
merge
Revision #3168.1.43 Thu 2012-01-26 19:25:22 +0100
Disable perfschema/all_instances for the threadpool (because of new mutexes and conditions)
merge] Thu 2012-01-26 17:35:01 +0100
merge
Revision #3168.1.41 Thu 2012-01-26 04:35:54 +0100
Further review points and simplify Windows implementation
Revision #3168.1.40 Wed 2012-01-25 03:59:09 +0100
add test thread_pool_min_basic
Revision #3168.1.39 Tue 2012-01-24 19:18:22 +0100
further reduce diffs to 5.5, monty review
Revision #3168.1.38 Tue 2012-01-24 03:23:14 +0100
small cleanups
Revision #3168.1.37 Tue 2012-01-24 02:26:29 +0100
reduce diffs to the 5.5 version, remove random change in mysql-test-run.pl
merge] Tue 2012-01-24 01:59:03 +0100
merge
Revision #3168.1.35 Wed 2012-01-18 21:12:04 +0100
ensure that lock is held, whenever active thread counter changes. It was not the case inside listener routine.
Revision #3168.1.34 Tue 2012-01-17 18:50:40 +0100
Threadpool : address some of the monty's review points Also, print message when pool blocks.
Revision #3168.1.33 Mon 2012-01-16 02:18:24 +0100
Fix threadpool on BSD and Solaris
Revision #3168.1.32 Sun 2012-01-15 15:41:25 +0100
Get rid of idle thread counter atomic variable.
Instead, use function that loops over groups and calculates idle threads for "show status".
Revision #3168.1.31 Sun 2012-01-15 11:17:45 +0100
Threadpool -address review comments
Revision #3168.1.30 Fri 2012-01-13 15:53:17 +0100
Simplify thread attach/detach. Use connection specific mysys_var, rather than sharing worker thread's my_thread_var with THD.
Revision #3168.1.29 Thu 2012-01-12 13:40:09 +0100
fix kill test, again
Revision #3168.1.28 [merge] Wed 2012-01-11 14:56:19 +0100
merge 5.5
Revision #3168.1.27 Tue 2012-01-10 16:58:30 +0100
MDEV-82 : Fix place in thr_lock.c where wait-end callback was called without corresponding wait-begin
Revision #3168.1.26 [merge] Mon 2012-01-02 12:03:49 +0100
merge with 5.5
Revision #3168.1.25 Mon 2012-01-02 11:43:22 +0100
Fix crashes in windows-embedded
Revision #3168.1.24 Mon 2012-01-02 10:13:53 +0100
fix test
Revision #3168.1.23 Sat 2011-12-31 05:24:11 +0100
fixes Bug #909774
Allow for faster creation of threads in corner cases where pool would be overloaded with long non-yielding queries.
To allow it, change minimum of thread_pool_stall_limit to be 10 milliseconds.
Also introduce a new parameter to oversubscribe a group . Number of threads running in parallel would be higher than it normally should, leading to thrashing, but it may improving preemptiveness, which is useful for the described corner case.
Revision #3168.1.22 Thu 2011-12-29 21:11:06 +0100
Make threadpool_stall_limit variable really dynamic
Revision #3168.1.21 Thu 2011-12-29 19:37:26 +0100
Bug #909537: Ensure thd_wait_begin/thd_wait_end callbacks are called.
Revision #3168.1.20 Thu 2011-12-29 13:37:37 +0100
Fix valgrind errors with network timeouts.
merge] Thu 2011-12-29 12:54:40 +0100
merge
Revision #3168.1.18 Thu 2011-12-29 12:53:07 +0100
Bug #909512: Fix crash on tp_set_threadpool_size if threadpool is not used(thread_handling != pool-of-threads)
Revision #3168.1.17 Thu 2011-12-29 12:17:30 +0100
Fix Bug #909414: Valgrind warnings in threadpool code
Revision #3168.1.16 Thu 2011-12-29 01:59:05 +0100
bug: 9091416: destroy timer mutex when threadpool scheduler shuts down.
Fixes valgrind warning.
Revision #3168.1.15 Wed 2011-12-28 16:23:46 +0100
use performance-schema friendly mysql_thread_create() instead of pthread_create()
Revision #3168.1.14 Wed 2011-12-28 03:51:12 +0100
fix result file
Revision #3168.1.13 Tue 2011-12-27 17:54:04 +0100
disable threadpool threads in sys_var suite, when the suite runs with embedded server
Revision #3168.1.12 Tue 2011-12-27 16:10:34 +0100
fix test suite
Revision #3168.1.11 Tue 2011-12-27 12:20:06 +0100
fix embedded build and warning
merge] Mon 2011-12-26 16:57:28 +0100
merge
Revision #3168.1.9 Mon 2011-12-26 01:08:46 +0100
Fix build on old 32 bit Centos (kernel 2.6.18)
Revision #3168.1.8 Wed 2011-12-21 00:56:34 +0100
Fix threadpool related test failures
Revision #3168.1.7 Tue 2011-12-20 22:49:24 +0100
make sys_vars suite pass
Revision #3168.1.6 Mon 2011-12-19 13:28:30 +0100
allow changing thread_pool_size without server restart
Revision #3168.1.5 Sun 2011-12-18 23:03:35 +0100
Fix pool_of_threads test case
Revision #3168.1.4 Sun 2011-12-18 20:40:38 +0100
Small adjustements to threadpool
Revision #3168.1.3 Wed 2011-12-14 23:16:50 +0100
On Unix, correct default threadpool_idle_timeout to be 60 sec
Revision #3168.1.2 Sat 2011-12-10 19:35:44 +0100
Fix Unix build
Revision #3168.1.1 Thu 2011-12-08 19:17:49 +0100
Initial threadpool implementation for MariaDB 5.5
Fix memory leak in one error case in mysqldump.
Fix that HAVE_VALGRIND_VALGRIND_H is now HAVE_VALGRIND in 5.5.
Fix that @have_ssl should not be set in embedded (introduced when removing #undef HAVE_OPENSSL from my_global.h).
Revision #3253.1.3 [merge] Wed 2012-02-22 12:14:34 +0100
Merge latest MariaDB 5.5 into MWL#192: Non-blocking client library.
Revision #3253.1.2 Wed 2012-02-22 11:44:53 +0100
Fix MWL#192 build error: Remove SSL special case for embedded server.
VIO has SSL in embedded server anyway, so we do not win anything by excluding it.
This was actually already done in this changeset:
revision-id: kostja@sun.com-20100413150445-8x23keoxdiufgq76
"... Also, he removed the (probable) bug of embedded server never using SSL-dependent functions..."
But was apparenly lost by a mis-merge of WL#5030.
Revision #3253.1.1 [merge] Tue 2012-02-21 22:15:44 +0100
Merge MWL#192: Non-blocking client library, into MariaDB 5.5.
Revision #2502.531.5 Mon 2012-01-30 13:45:58 +0100
MWL#192: Fix problem when we first enable MYSQL_OPT_NONBLOCK, then connect in normal blocking style, then later do a non-blocking operation.
In this case, the vio->async_context was not set up correctly, so that non-blocking operation was not properly handled.
Sun 2012-01-08 10:13:39 +0100
Fix non-ssl build.
Sat 2012-01-07 23:20:08 +0100
fix typo.
Fri 2012-01-06 12:43:18 +0100
: non-blocking client API, after-review fixes.
Main change is that non-blocking operation is now an option that must be explicitly enabled with mysql_option(mysql, MYSQL_OPT_NONBLOCK, ...) before any non-blocing operation can be used.
Also the CLIENT_REMEMBER_OPTIONS flag is now always enabled and thus effectively ignored (it was not really useful anyway, and this simplifies things when non-blocking mysql_real_connect() fails).
Tue 2011-09-20 12:49:25 +0200
: Non-blocking client API for libmysqlclient.
All client functions that can block on I/O have alternate _start() and _cont() versions that do not block but return control back to the application, which can then issue I/O wait in its own fashion and later call back into the library to continue the operation.
Works behind the scenes by spawning a co-routine/fiber to run the blocking operation and suspend it while waiting for I/O. This co-routine/fiber use is invisible to applications.
The bug was fixed by the patch for Bug #914560.
Revision #2502.538.5 Tue 2012-01-17 03:26:49 -0800
Fixed Bug #914560.
The patch for MWL#247 forgot to initialize the TABLE::ext_key_parts and TABLE::ext_key_flags of the temporary tables by a query. This could cause crashes for queries the execution of which needed creation of temporary tables.
Revision #2502.538.4 Sun 2012-01-01 22:42:11 -0800
Fixed more compiler warnings.
Revision #2502.538.3 Sun 2012-01-01 21:41:57 -0800
Fixed compiler warnings.
Revision #2502.538.2 Sun 2012-01-01 20:47:36 -0800
Fixed calculation of rec_per_key elements for added components of the extended keys.
Slightly corrected the implementation of the function ha_innobase::read_time().
Changed the implementation of handler::keyread_time to make the cost of single key index only look-ups dependent on the key entry length.
Corrected the index of the last possible components of an extended key in the function best_access_path().
Revision #2502.538.1 [merge] Sat 2011-12-31 03:36:20 -0800
Merged MWL#247 into the latest 5.3.
Revision #2502.537.1 Sat 2011-12-31 02:25:57 -0800
Implementation of the MWL#247: Make the optimizer use extended keys.
The main patch.
For i368/x86_64 on GCC, uses very fast assembler co-routine support. On Windows uses native Win32 Fibers. Falls back to POSIX ucontext on other platforms. Assembler routines for more platforms are relatively easy to add by extending mysys/my_context.c, eg. similar to the Lua lcoco library.
For testing, mysqltest and mysql_client_test are extended with the
option --non-blocking-api. This causes the programs to use the
non-blocking API for database access. mysql-test-run.pl has a similar
option --non-blocking-api that uses this, as well as additional
testcases.
An example program tests/async_queries.c is included that uses the new non-blocking API with libevent to show how, in a single-threaded program, to issue many queries in parallel against a database.
This page is licensed: CC BY-SA / Gnu FDL