maria-developers team mailing list archive

[Sergey Petrunya <psergey@xxxxxxxxxxxx>]

Hi!

I've got this test failure in maria-5.1-table-elimination tree:

main.mysql-bug41486                      [ fail ]

http://askmonty.org/buildbot/builders/jaunty-amd64-rel/builds/54/steps/test/logs/stdio

CURRENT_TEST: main.mysql-bug41486
--- .../r/mysql-bug41486.result
+++ .../r/mysql-bug41486.reject
@@ -8,6 +8,5 @@
 SET @@global.general_log = @old_general_log;
 SELECT LENGTH(data) FROM t1;
 LENGTH(data)
-2097152
 DROP TABLE t1;
 SET @@global.max_allowed_packet = @old_max_allowed_packet;

mysqltest: Result length mismatch


Here's the relevant part of the .test file:

  CREATE TABLE t1(data LONGBLOB);
  INSERT INTO t1 SELECT REPEAT('1', 2*1024*1024);

  let $outfile= $MYSQLTEST_VARDIR/tmp/bug41486.sql;
  --error 0,1
  remove_file $outfile;
  --exec $MYSQL_DUMP test t1 > $outfile
  SET @old_general_log = @@global.general_log;
  SET @@global.general_log = 0;
  # Check that the mysql client does not insert extra newlines when loading
  # strings longer than client's max_allowed_packet
  --exec $MYSQL --max_allowed_packet=1M test < $outfile 2>&1
  SET @@global.general_log = @old_general_log;
  SELECT LENGTH(data) FROM t1;

My analysis relvealed that this part of the test

  INSERT INTO t1 SELECT REPEAT('1', 2*1024*1024);

  let $outfile= $MYSQLTEST_VARDIR/tmp/bug41486.sql;
  --error 0,1
  remove_file $outfile;
  --exec $MYSQL_DUMP test t1 > $outfile

gets executed as follows: when $MYSQL_DUMP runs the 

  SELECT /*!40001 SQL_NO_CACHE */ * FROM `t1`

statement to get the table data, the select produces nothing, even though
INSERT statement has already finished by that time (at least from client 
point of view).

The reason for select producing nothing is that the optimizer identifies
table t1 as constant (it has one or zero rows), then it tries to get the
record with handler->read_first_row() call, and it gets HA_ERR_END_OF_FILE.

So far I've fixed the test case by adding SELECT COUNT(*) FROM t1 (as an
arbitrary select statement involving t1) after the INSERT.

The questions are:
- Is the above behavior expected of MyISAM? (I suppose it is but I'm not
  sure)
- Any ideas why does this suddenly show up when I make totally unrelated
  changes in table elimination code. The changed part of the code is never
  executed by the test...

BR
 Sergey
-- 
Sergey Petrunia, Software Developer
Monty Program AB, http://askmonty.org
Blog: http://s.petrunia.net/blog