Hi Igor, Timour, Sergey, Can one of you please check this patch from MySQL 5.1.46? It is a commit to fix http://bugs.mysql.com/bug.php?id=51494 This patch introduces a regression: http://bugs.mysql.com/bug.php?id=53334 If I revert the patch, the regression disappears. And interestingly, the included test case does not fail even when the fix is reverted (go figure...) So it would be good if one of you could check the patch and check what is wrong with it, and if a different fix for Bug#51494 is needed. (needed to complete merge of MySQL 5.1.46). Thanks, - Kristian. ------------------------------------------------------------ revno: 3407.1.1 revision-id: sergey.glukhov@sun.com-20100319060102-57ykzjf4pc93avy1 parent: omer@mysql.com-20100318064207-l3ap0mpxt510b4n3 committer: Sergey Glukhov <Sergey.Glukhov@sun.com> branch nick: mysql-5.1-bugteam timestamp: Fri 2010-03-19 10:01:02 +0400 message: Bug#51494 crash with join, explain and 'sounds like' operator The crash happens because of discrepancy between values of conts_tables and join->const_table_map(make_join_statisctics). Calculation of conts_tables used condition with HA_STATS_RECORDS_IS_EXACT flag check. Calculation of join->const_table_map does not use this flag check. In case of MERGE table without union with index the table does not become const table and thus join_read_const_table() is not called for the table. join->const_table_map supposes this table is const and later in make_join_select this table is used for making&calculation const condition. As table record buffer is not populated it leads to crash. The fix is adding a check if an engine supports HA_STATS_RECORDS_IS_EXACT flag before updating join->const_table_map. diff: === modified file 'sql/sql_select.cc' --- sql/sql_select.cc 2010-03-14 16:01:45 +0000 +++ sql/sql_select.cc 2010-03-19 06:01:02 +0000 @@ -2943,7 +2943,8 @@ s->quick=select->quick; s->needed_reg=select->needed_reg; select->quick=0; - if (records == 0 && s->table->reginfo.impossible_range) + if (records == 0 && s->table->reginfo.impossible_range && + (s->table->file->ha_table_flags() & HA_STATS_RECORDS_IS_EXACT)) { /* Impossible WHERE or ON expression === modified file 'mysql-test/r/merge.result' --- mysql-test/r/merge.result 2010-03-03 10:49:03 +0000 +++ mysql-test/r/merge.result 2010-03-19 06:01:02 +0000 @@ -2286,4 +2286,16 @@ DROP TABLE m1; DROP TABLE `test@1`.`t@1`; DROP DATABASE `test@1`; +# +# Bug#51494c rash with join, explain and 'sounds like' operator +# +CREATE TABLE t1 (a INT) ENGINE=MYISAM; +INSERT INTO t1 VALUES(1); +CREATE TABLE t2 (b INT NOT NULL,c INT,d INT,e BLOB NOT NULL, +KEY idx0 (d, c)) ENGINE=MERGE; +EXPLAIN SELECT * FROM t1 NATURAL RIGHT JOIN +t2 WHERE b SOUNDS LIKE e AND d = 1; +id select_type table type possible_keys key key_len ref rows Extra +1 SIMPLE NULL NULL NULL NULL NULL NULL NULL Impossible WHERE noticed after reading const tables +DROP TABLE t2, t1; End of 5.1 tests === modified file 'mysql-test/t/merge.test' --- mysql-test/t/merge.test 2010-03-03 10:49:03 +0000 +++ mysql-test/t/merge.test 2010-03-19 06:01:02 +0000 @@ -1690,4 +1690,19 @@ DROP TABLE `test@1`.`t@1`; DROP DATABASE `test@1`; +--echo # +--echo # Bug#51494c rash with join, explain and 'sounds like' operator +--echo # + +CREATE TABLE t1 (a INT) ENGINE=MYISAM; +INSERT INTO t1 VALUES(1); + +CREATE TABLE t2 (b INT NOT NULL,c INT,d INT,e BLOB NOT NULL, +KEY idx0 (d, c)) ENGINE=MERGE; + +EXPLAIN SELECT * FROM t1 NATURAL RIGHT JOIN +t2 WHERE b SOUNDS LIKE e AND d = 1; + +DROP TABLE t2, t1; + --echo End of 5.1 tests