Bug #152 wrong results for latin1_de
Submitted: 14 Mar 2003 15:49 Modified: 27 Apr 2003 6:49
Reporter: Sergei Golubchik Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server Severity:S3 (Non-critical)
Version:4.0 OS:
Assigned to: Alexander Barkov CPU Architecture:Any

[14 Mar 2003 15:49] Sergei Golubchik
Description:
wrong results for latin1_de

How to repeat:
drop table if exists t1;
create table t1 (word varchar(255) not null default '', index(word));
insert into t1 values ('ss'),('"'),('D'),('ae');
select * from t1 where word='ss'; -- ok
select * from t1 where word='"'; -- bug
select * from t1 where word like 'ae'; -- bug
select * from t1 where word like 'D'; -- bug
drop table t1;
[14 Mar 2003 16:01] Sergei Golubchik
grr.

insert into t1 values ('ss'),(0xDF),(0xE4),('ae');

" was supposed to be ß
D - ä
[20 Apr 2003 7:24] Alexandre Vaniachine
The latin1_de test should be modified to include the
require file that will check for the presense of the extended char set.
Right now this test hangs the mysql-test-run for the servers compiled
without the extended char set support.
[27 Apr 2003 6:49] Michael Widenius
Thank you for your bug report. This issue has been fixed in the latest
development tree for that product. You can find more information about
accessing our development trees at 
    http://www.mysql.com/doc/en/Installing_source_tree.html

Fixed in the 4.0 tree.

Note that LIKE comparision is done character by character;  This will probably be fixed in the 4.1 tree.