Hi Alfranio,
Patch is approved with some items I would like you to fix. I also have
some minor comments below.
* Please make it clear in the comments and the patch that
IMMEDIATE_LOGGING implies that the statement is written without
BEGIN/COMMIT. Probably best to add to the comment to
binlog_start_trans_and_stmt().
* Normally, the error codes should not change and it seems like
they did. If different errors are generated, that can be
acceptable, but error codes should not change if the same error
is generated.
Just my few cents,
Mats Kindahl
On 03/21/2011 04:56 PM, Alfranio Correia wrote:
> === modified file 'mysql-test/r/read_only_innodb.result'
> --- a/mysql-test/r/read_only_innodb.result 2011-03-16 16:41:46 +0000
> +++ b/mysql-test/r/read_only_innodb.result 2011-03-21 15:56:26 +0000
> @@ -97,7 +97,7 @@ a
> 2
> ROLLBACK;
> Warnings:
> -Warning 1726 The creation of some temporary tables could not be rolled back.
> +Warning 1732 The creation of some temporary tables could not be rolled back.
Why did the error code for the warning message change? It is usually
an error if the code of an error changes since there are application
that rely on the error code to be identical between versions.
> === modified file 'mysql-test/suite/rpl/r/rpl_row_binlog_max_cache_size.result'
> --- a/mysql-test/suite/rpl/r/rpl_row_binlog_max_cache_size.result 2011-02-23 20:01:27
> +0000
> +++ b/mysql-test/suite/rpl/r/rpl_row_binlog_max_cache_size.result 2011-03-21 15:56:26
> +0000
> @@ -150,7 +150,7 @@ SET GLOBAL binlog_stmt_cache_size= ORIGI
> BEGIN;
> Repeat statement 'INSERT INTO t1 VALUES($n, repeat("a", 32))' 128 times
> COMMIT;
> -include/wait_for_slave_sql_error.inc [errno=1534]
> +include/wait_for_slave_sql_error.inc [errno=1197]
Why does the error code change? Is it really a different error or has
the error code just changed?
> @@ -121,6 +129,11 @@ public:
> return(incident);
> }
>
> + bool is_trx_cache()
> + {
> + return trx_cache;
> + }
> +
Constant member function?
