- 08 Jul, 2013 11 commits
-
-
Sergey Petrunya authored
- After the merge from mysql-5.6, open_tables() did not call open_and_process_table() for temporary table. The logic was that temporary tables were already opened when mysql_execute_command() has called open_temporary_tables(). This worked for the most part, except for temporary tables of type MERGE. for which open_and_process_table() must call table->file->extra(HA_EXTRA_ADD_CHILDREN_LIST). Failure to make this call resulted in crash further in execution. tables->table->file->extra(HA_EXTRA_ADD_CHILDREN_LIST)
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
unknown authored
-
Sergey Petrunya authored
- Put back the code tht produces the warning about "Table storage engine %s does not support the create option 'TRANSACTIONAL=1'"
-
Sergey Petrunya authored
-
unknown authored
Merged the majority of the PFS test cases. There are still 19 failing test cases that need more attention.
-
Sergey Petrunya authored
-Note 1031 Table storage engine for 't1' doesn't have this option +Note 1031 Table storage engine for 'InnoDB' doesn't have this option They were caused by a change in MariaDB which changed ER_ILLEGAL_HA message text to be like: "Storage engine InnoDB of the table `test`.`t1` doesn't have this option" Some the error calls were changed to pass new parameters, but some were left to be old. Also the error text in errmsg-ut8.txt was not changed.
-
Sergey Petrunya authored
- Update test results for tests using SPs: SPs no longer emit warnings/errors that were caught and handled inside SP
-
Sergey Petrunya authored
-
unknown authored
-
- 07 Jul, 2013 1 commit
-
-
Sergey Petrunya authored
- New error message text - PARTITION is now a reserved word in SQL, so it should be quoted
-
- 05 Jul, 2013 20 commits
-
-
Sergey Petrunya authored
-
Sergey Petrunya authored
- Update trivial .reject/.result differences (all checked)
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
Sergey Petrunya authored
must use C++ linking due to client library using SSL library, which needs C++ linking
-
Sergey Petrunya authored
- Take into account that Dynamic_array::back() now returns pointer to the last element (it used to return pointer to right after the last element) - Fix error messages merge: ER_INTERNAL_ERROR was defined independently by both mysql-5.6 and mariadb-5.5. Switch to their error number, and still support ours for compatibility.
-
Sergey Petrunya authored
-
unknown authored
-
unknown authored
-
unknown authored
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
Vladislav Vaintroub authored
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
Sergey Petrunya authored
- strict.test updated (changed back) after the cset with "Fix type_newdecimal.test ..." two csets ago - row-checksum.test changed the code from HA_WRONG_CREATE_OPTION to ER_ILLEGAL_HA_CREATE_OPTION, like mysql-5.6 did
-
Sergey Petrunya authored
- MariaDB does not have mysql.slave_master_info, mysql.slave_relay_log_info ,mysql.slave_worker_info or mysql.ndb_binlog_index tables. - Some tests expected to have these tables (this was an incorrect merge from 5.6, which merged necessary tables like mysql.innodb*stats, but also got these tables)
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
- 04 Jul, 2013 8 commits
-
-
Sergey Petrunya authored
-
Sergey Petrunya authored
Update log_tables.test: some definitions of columns have changed, and slow_log got 'thread_id' column.
-
Sergey Petrunya authored
-
Sergey Petrunya authored
Fix fix_priv_tables.test: make mysql_system_tables_fix.sql to not modify user.password_expired column.
-
Sergey Petrunya authored
-
Sergey Petrunya authored
-
Sergey Petrunya authored
- make the test output stable - make Log_to_csv_event_handler::log_slow() to write the value of thd->thread_id (it didn't, and so 0 was always logged).
-
Sergey Petrunya authored
-