- 03 Mar, 2011 1 commit
-
-
Tor Didriksen authored
This was a buffer overrun in do_div_mod(), overwriting the internal buffer of auto variable 'tmp' in Item_func_int_div::val_int. Result on windows: 'this' is set to zero, and crash. Ran fine on other platforms (no valgrind warnings), but this is undefined behaviour on any platform of course.
-
- 17 Mar, 2011 9 commits
-
-
Georgi Kodinov authored
-
Kent Boortz authored
-
Georgi Kodinov authored
Added the cleanup code. Extended mysql_client_test to take --plugin-dir and --default_auth. Added a C test case.
-
Alfranio Correia authored
-
Nirbhay Choubey authored
with no .cnf or .ini extension. Fix for this bug was pushed as part of Bug#11765482.
-
Alfranio Correia authored
-
Nirbhay Choubey authored
-
Nirbhay Choubey authored
-
Sven Sandberg authored
Problem: the test had not been updated after BUG#49978 was pushed Fix: add 'set $rpl_only_running_threads= 1' to the end of the test and update the result file. Also, use include/assert.inc for an assertion (instead of relying on result file comparison). Also, move 'set @@global.slave_net_timeout' forwards, to get rid of a warning.
-
- 16 Mar, 2011 7 commits
-
-
Kent Boortz authored
-
Kent Boortz authored
-
Kent Boortz authored
-
Alfranio Correia authored
Currently, rpl_semi_sync is failing in PB due to the warning message: "Slave SQL: slave SQL thread is being stopped in the middle of " "applying of a group having updated a non-transaction table; " "waiting for the group completion ..." The problem started happening after the fix for BUG#11762407 what was automatically suppressing some warning messages. To fix the current issue, we suppress the aforementioned warning message and exploit the opportunity to make the sentence clearer.
-
Luis Soares authored
Automerge: mysql-5.1 --> mysql-5.5
-
Luis Soares authored
There is a race between two threads: user thread and the dump thread. The former sets a debug instruction that makes the latter wait before processing an Xid event. There can be cases that the dump thread has not yet processed the previous Xid event, causing it to wait one Xid event too soon, thus causing sync_slave_with_master never to resume. We fix this by moving the instructions that set the debug variable after calling sync_slave_with_master.
-
hery.ramilison@oracle.com authored
-
- 15 Mar, 2011 6 commits
-
-
Serge Kozlov authored
-
Sven Sandberg authored
Problem: the test failed because errors were found in the error log. The test case contains suppressions for an old version of the error message, but the format of the error message has changed without updating the suppression. Fix: Update the suppression. Also small fixes to improve the test.
-
Dmitry Shulga authored
of service in prepared statements).
-
Dmitry Shulga authored
The problem was that server didn't check resulting size of prepared statement argument which was set using mysql_send_long_data() API. By calling mysql_send_long_data() several times it was possible to create overly big string and thus force server to allocate memory for it. There was no way to limit this allocation. The solution is to add check for size of result string against value of max_long_data_size start-up parameter. When intermediate string exceeds max_long_data_size value an appropriate error message is emitted. We can't use existing max_allowed_packet parameter for this purpose since its value is limited by 1GB and therefore using it as a limit for data set through mysql_send_long_data() API would have been an incompatible change. Newly introduced max_long_data_size parameter gets value from max_allowed_packet parameter unless its value is specified explicitly. This new parameter is marked as deprecated and will be eventually replaced by max_allowed_packet parameter. Value of max_long_data_size parameter can be set only at server startup.
-
Jon Olav Hauglid authored
FAILED DROP DATABASE CAN BREAK STATEMENT BASED REPLICATION The first phase of DROP DATABASE is to delete the tables in the database. If deletion of one or more of the tables fail (e.g. due to a FOREIGN KEY constraint), DROP DATABASE will be aborted. However, some tables could still have been deleted. The problem was that nothing would be written to the binary log in this case, so any slaves would not delete these tables. Therefore the master and the slaves would get out of sync. This patch fixes the problem by making sure that DROP TABLE is written to the binary log for the tables that were in fact deleted by the failed DROP DATABASE statement. Test case added to binlog.binlog_database.test.
-
Bjorn Munch authored
-
- 14 Mar, 2011 5 commits
-
-
Chuck Bell authored
-
Davi Arnaut authored
-
Davi Arnaut authored
Add a missing DBUG_RETURN function test_if_number().
-
Magne Mahre authored
INSTALLATION When starting mysqld as an MS Windows NT service, it crashed with "Error 1067: The process terminated unexpectedly". The problem is that thread local variables are not allocated and initialized properly when started as a service. When the server is started as a regular executable, the problem does not occur. Analysis showed that this is a regression after the patch for Bug#11765237/Bug#11763065. Before, the thread local storage was initialized by the call chain: win_main->my_basic_init->my_thread_basic_global_init-> my_thread_init When the my_init() structure was changed, this initialization was moved from win_main to mysqld_main. When started as a service win_main is run in a separate thread, which does not have mysqld_main in its call path, so my_thread_init is never called for this thread. Added a call to my_thread_init / my_thread_end in the service handler function, which solves the problem.
-
Alexander Nozdrin authored
fails when running with ps-protocol). The problem was that when running in --ps-protocol mode mysqltest.cc didn't close created prepared statements. So, the plugins could not be unistalled because there was a prepared statement using them. A fix is to add a dummy statement that forces mysqltest.cc to close the last prepared statement (which uses a plugin-defined table).
-
- 11 Mar, 2011 11 commits
-
-
Chuck Bell authored
This patch corrects the problem by fixing the definition and alterations of the mysql.user table in the .sql files. Also included are new result files for tests that examine the name column of the mysql.user table.
-
Joerg Bruehe authored
When a RPM test build in a non-release branch is done, the $MYSQL_BINDIR variable ends in "/usr" (rather than in "/usr/lib" as in a RPM release build), this made test "file_contents" fail. A branch for this case is added to the test. The test result is unchanged.
-
Bjorn Munch authored
-
Bjorn Munch authored
-
Bjorn Munch authored
-
Mayank Prasad authored
-
Mayank Prasad authored
Issue: SSL_CIPHER set to a specific CIPHER name was not getting picked up by SHOW STATUS Command. Solution: If specific cipher name is specified, avoid overwriting of Cipher List with default Cipher names.
-
Marc Alff authored
-
Bjorn Munch authored
-
Bjorn Munch authored
-
Bjorn Munch authored
-
- 10 Mar, 2011 1 commit
-
-
Marc Alff authored
-