mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2026-05-16 03:47:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

Fixed bug#36005: server crashes inside NOT IN clause subquery with

Browse source 
impossible WHERE/HAVING clause
                 (subselect_single_select_engine::exec).

Allocation and initialization of joined table list t1, t2... of
subqueries like:

    NOT IN (SELECT ... FROM t1,t2,... WHERE 0)

is optimized out, however server tries to traverse this list.


mysql-test/r/subselect3.result:
  Added test case for bug#36005.
mysql-test/t/subselect3.test:
  Added test case for bug#36005.
sql/sql_select.cc:
  Fixed bug#36005.
  
  1. JOIN::prepare initializes JOIN::table counter (actually a size
     of the JOIN::join_tab array) and sets it to a number of joined tables.
  
  2. The make_join_statistics function (when called from JOIN::optimize)
     allocates and fills the JOIN::join_tab array.
     However, when optimizing subselect has impossible (definite false)
     WHERE or HAVING clause, optimizer skips call to make_join_statistics
     and leaves JOIN::join_tab == NULL.
  
  3. subselect_single_select_engine::exec does traversal of the JOIN::join_tab
     array and the server dies because array is not allocated but array
     counter is greater than 0.
  
  The JOIN::optimize method has been modified to reset the JOIN::table
  counter to 0 in cause of impossible WHERE/HAVING clause.
This commit is contained in:
unknown 2008-04-23 02:27:23 +05:00
commit 73f7de59c2
3 changed files with 23 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

9
mysql-test/r/subselect3.result
View file

@ -770,4 +770,13 @@ SELECT ROW(1, 2) IN (SELECT t1.a, 2 FROM t2) FROM t1 GROUP BY t1.a;
ROW(1, 2) IN (SELECT t1.a, 2 FROM t2)
1
DROP TABLE t1, t2;
CREATE TABLE t1 (a INT);
INSERT INTO t1 VALUES (1),(2),(3);
CREATE TABLE t2 SELECT * FROM t1;
SELECT 1 FROM t1 WHERE t1.a NOT IN (SELECT 1 FROM t1, t2 WHERE 0);
1
1
1
1
DROP TABLE t1, t2;
End of 5.0 tests