From: Roy Lyseng
Date: April 11 2011 9:13am
Subject: Re: bzr commit into mysql-trunk branch (roy.lyseng:3358) Bug#11822517
List-Archive: http://lists.mysql.com/commits/135168
Message-Id: <4DA2C62A.6050900@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit

On 11.04.11 10.05, Øystein Grøvlen wrote:


>>> > + static uint determine_search_depth(JOIN* join, uint search_depth);
>>>
>>> I assume this is static in order to be able to call it from the
>>> initialization list. I think I would prefer, making this a non-static
>>> member function, remove the join parameter, and moving the
>>> initialization to the body of the constructor.
>>
>> It means that search_depth will no longer be const in the class. Is that
>> Ok?
>
> Good point.
> It is OK with me, but I will leave it to you to decide.

I think I keep it static. Copying it to the new class becomes pretty meaningless 
if it is modifiable, in particular when we have the argument 
current_search_depth supplied to the member functions as well.

I made the determine_search_depth() function independent of join, btw, by adding 
table_count as argument.

>>> > + DBUG_ASSERT((join_tables& join->const_table_map) == 0);
>>>
>>> I do not think this assert is meaningful. Given the assignement
>>> above, it should always be true.
>>
>> It will become meaningful when the second patch is added ;)
>
> I think that means that the assert really belongs to the second patch.

Of course :)

Thanks,
Roy