From: Jorgen Loland
Date: June 20 2011 10:47am
Subject: Re: bzr commit into mysql-trunk branch (guilhem.bichot:3385) Bug#12586926
 Bug#12619510 Bug#12619868
List-Archive: http://lists.mysql.com/commits/139497
Message-Id: <4DFF2524.1060004@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit

Hi Guilhem,

The patch is approved but see suggestion inline.

On 06/17/2011 02:46 PM, Guilhem Bichot wrote:
>          4) In testcase of BUG#12619868, we had two wrong extra rows "NULL,4" and "NULL,8".
>          - 1	SIMPLE	alias2	index	NULL	col_int_key	4	NULL	15	
>          - 1	SIMPLE	alias2	ALL	NULL	NULL	NULL	NULL	16	Using temporary; Using filesort

+ on the first line?

>          Before the bufxig: as earlier, JOIN::sort_by_table is NULL, because when

bugfix

> === modified file 'sql/sql_select.cc'
> --- a/sql/sql_select.cc	2011-06-11 13:38:32 +0000
> +++ b/sql/sql_select.cc	2011-06-17 12:45:22 +0000
> @@ -2328,24 +2327,26 @@ JOIN::optimize()
>         sort_and_group= 0;
>     }
>
> -  // Can't use sort on head table if using join buffering
> -  if (full_join)
> +  select_opts_for_readinfo=
> +    (select_options&  (SELECT_DESCRIBE | SELECT_NO_JOIN_CACHE)) |
> +    (select_lex->ftfunc_list->elements ?  SELECT_NO_JOIN_CACHE : 0);
> +

Please move initialization of select_opts_for_readinfo immediately before you 
need it. You may also move it's declaration down here, but that's a question of 
preference.

-- 
Jørgen Løland | Senior Software Engineer | +47 73842138
Oracle MySQL
Trondheim, Norway