| Bug #26940 | CREATE SERVER does not initialize record[0] in insert_server_record() | ||
|---|---|---|---|
| Submitted: | 7 Mar 2007 23:42 | Modified: | 8 Mar 2007 21:17 |
| Reporter: | Timothy Smith | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server: Tests | Severity: | S7 (Test Cases) |
| Version: | OS: | ||
| Assigned to: | Magnus Blåudd | CPU Architecture: | Any |
[8 Mar 2007 1:00]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/21440 ChangeSet@1.2460, 2007-03-08 02:00:21+01:00, tsmith@quadxeon.mysql.com +1 -0 Bug #26940: CREATE SERVER does not initialize record[0] in insert_server_record() sql_servers.cc:insert_server_record(): initialize record[0] before inserting row
[8 Mar 2007 21:17]
Paul DuBois
No changelog entry needed.
Description: Quote from Magnus: -------------------------- while working with unifying the sources for our system tables(they are all defined in scripts/mysql_system_tables.sql now) I found that we inserted a "test" server in mysql.server, since we don't really want to have any test data in the official MySQL system tables I removed that and it reveals that CREATE SERVER does not properly initialize record[0] before inserting the record Using 'empty_record' in that function would fix it. --- 1.5/sql/sql_servers.cc 2007-01-29 00:47:31 +01:00 +++ edited/sql_servers.cc 2007-03-05 16:42:27 +01:00 @@ -545,6 +545,8 @@ int insert_server_record(TABLE *table, F DBUG_ENTER("insert_server_record"); table->use_all_columns(); + empty_record(table); + /* set the field that's the PK to the value we're looking for */ table->field[0]->store(server->server_name, server->server_name_length, What do you say about that? At first it looked like this only occurred on Solaris and Windows but it can actually be reproduced on a normal linux machine if removing the code that initializes memory to 0xA5. A little funny that our code to find bugs actually masks it. -------------------------- Timothy added the bogus row into the mysql.servers table as a stop-gap measure until this patch can be approved. Please remove that from scripts/msyql_system_tables_data.sql once this is approved. How to repeat: See, for example: https://intranet.mysql.com/~knielsen/pb/getlog.pl?dir=mysql-5.1-new-maint&entry=tsmith@qua... Suggested fix: See description.