He Zhenxing Wrote:
> Hi Guangbao,
>
> Thank you for your work! Here are some comments:
>
> The commit message for a bug usually includes three parts:
> 1) the bug number and the title
> 2) the reason that caused the problem
> 3) your fix of the problem
>
> Please add 2) and 3) to your commit message of this patch.
>
Sorry, I think the title and the patch of the bug could show enough
info. :-)
I will add them.
> And please also add test for this bug so that it will make sure this
> problem will not regression.
>
Okay, Will add.
> I also noticed in the bug report that about the issue when SQL_MODE is
> STRICT_TRANS_TABLES, non-trans tables such as MyISAM is also converting
> warnings into errors. Have you checked this issue?
>
Yes, I noticed that.
The crashing bug#42640 interrupts me to check it further deeply.
But I think probably the transaction.stmt.modified_non_trans_table
doesn't set correctly for (insert ...select ... clause),
which elevates warning to error.
Maybe it's a mysql server bug and unrelated to binary log. I am
investigating it more further now.
> Guangbao Ni wrote:
>
>> #At file:///home/ngb/mysql/bzr/bugteam-5.1-bug41980/
>>
>> 2793 Guangbao Ni 2009-02-11
>> BUG#41980 SBL, INSERT .. SELECT .. LIMIT = ERROR, even when @@SQL_LOG_BIN
> is 0 !
>> modified:
>> sql/sql_class.cc
>>
>> per-file messages:
>> sql/sql_class.cc
>> Add a condition (sql_bin_log=on) for if statement. then print a warning when
> it is in statement mode and trying to log an unsafe statement.
>> === modified file 'sql/sql_class.cc'
>> --- a/sql/sql_class.cc 2009-02-06 09:53:20 +0000
>> +++ b/sql/sql_class.cc 2009-02-11 19:12:57 +0000
>> @@ -3657,7 +3657,8 @@ int THD::binlog_query(THD::enum_binlog_q
>> If we are in statement mode and trying to log an unsafe statement,
>> we should print a warning.
>> */
>> - if (lex->is_stmt_unsafe() &&
>> + if ((options & OPTION_BIN_LOG) &&
>> + lex->is_stmt_unsafe() &&
>> variables.binlog_format == BINLOG_FORMAT_STMT)
>> {
>> DBUG_ASSERT(this->query != NULL);
>>
>>
>>
>
>
