From: Jorgen Loland
Date: August 26 2010 11:19am
Subject: Re: bzr commit into mysql-next-mr-bugfixing branch (jorgen.loland:3220)
 Bug#30597
List-Archive: http://lists.mysql.com/commits/116853
Message-Id: <4C764DDA.2070007@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit


On 08/25/2010 11:45 AM, Guilhem Bichot wrote:
> Hello,
>
> Jorgen Loland a écrit, Le 18.08.2010 11:01:
>> #At file:///export/home/jl208045/mysql/mysql-next-mr-bugfixing-30597/
>> based on revid:joerg@stripped
>>
>> 3220 Jorgen Loland 2010-08-18
>> Bug#30597: Change EXPLAIN output to include extrema of UNION components
>
>> === modified file 'sql/sql_lex.h'
>> --- a/sql/sql_lex.h 2010-08-16 18:21:24 +0000
>> +++ b/sql/sql_lex.h 2010-08-18 09:01:38 +0000
>> @@ -712,6 +712,13 @@ public:
>> }
>> st_select_lex* outer_select();
>> st_select_lex* next_select() { return (st_select_lex*) next; }
>> +
>> + inline st_select_lex* last_select() {
>
> you can declare this function "const", it doesn't change data.

If this object does not have a next, we effectively return 'this', which is 
const inside a const function. But we don't want the return pointer to be const, 
so I don't think that it makes sense to do this.

Simply changing to "inline st_select_lex* last_select() const {...}" gives 
compiler warning:

sql/sql_lex.h:765: error: invalid conversion from ‘const st_select_lex* const’ 
to ‘st_select_lex*’

Ok to push without const, or do you have another idea?

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