Details
Minor
Description
Consider the following terminal session demonstrating the bug:
MariaDB [(none)]> create database d1;
Query OK, 1 row affected (0.00 sec)
MariaDB [(none)]> use d1;
Database changed
MariaDB [d1]> create table t1 (e enum('one', 'two'));
Query OK, 0 rows affected (0.75 sec)
MariaDB [d1]> insert into t1 values ('one');
Query OK, 1 row affected (0.40 sec)
MariaDB [d1]> create table t2 (e enum('one', 'two')) as select * from t1;
Query OK, 1 row affected (0.78 sec)
Records: 1 Duplicates: 0 Warnings: 0
MariaDB [d1]> select * from t2;
+------+
| e |
+------+
| one |
+------+
1 row in set (0.01 sec)
MariaDB [d1]> select * from t2 union all select * from t2;
+------+
| e |
+------+
| o |
| o |
+------+
2 rows in set (0.00 sec)
MariaDB [d1]> alter table t2 engine MyISAM;
Query OK, 1 row affected (0.91 sec)
Records: 1 Duplicates: 0 Warnings: 0
MariaDB [d1]> select * from t2 union all select * from t2;
+------+
| e |
+------+
| one |
| one |
+------+
2 rows in set (0.00 sec)
MariaDB [d1]>
This bug exists in MySQL 5.1 too. It seems that the following one-liner fixes the problem but I'm not sure if it can break something else.
--- a/sql/sql_table.cc
+++ b/sql/sql_table.cc
@@ -3144,7 +3144,7 @@ mysql_prepare_create_table(THD *thd, HA_CREATE_INFO *create_info,
sql_field->charset= (dup_field->charset ?
dup_field->charset :
create_info->default_table_charset);
- sql_field->length= dup_field->char_length;
+ sql_field->length= dup_field->length;
sql_field->pack_length= dup_field->pack_length;
sql_field->key_length= dup_field->key_length;
sql_field->decimals= dup_field->decimals;
Gliffy Diagrams
Attachments
Issue Links
- links to
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
(The priority has been decreased because it's an upstream bug, reproducible on all of MySQL 5.1-5.6).
Hi Pavel,
Since you mentioned that it was reproducible on MySQL – have you already filed it for MySQL, or are you planning to?
If you have, could you please provide the bug number?
If you haven't and aren't planning to, we can do it on your behalf.
I haven't filed and am not planning to. Feel free to file yourself.
It looks like http://bugs.mysql.com/bug.php?id=14139 is relevant and the actual fix should look differently.