Hi Igor,
ibabaev@stripped wrote:
> Hi Alexey,
>
> See my notes below.
>
> Regards,
> Igor.
>
> Quoting Alexey Kopytov <alexeyk@stripped>:
>
>> diff -Nrup a/sql/item_sum.cc b/sql/item_sum.cc
>> --- a/sql/item_sum.cc 2007-12-21 13:44:21 +03:00
>> +++ b/sql/item_sum.cc 2008-02-21 20:24:58 +03:00
>> @@ -597,10 +597,7 @@ Item_sum_hybrid::fix_fields(THD *thd, It
>> result_field=0;
>> null_value=1;
>> fix_length_and_dec();
>> - if (item->type() == Item::FIELD_ITEM)
>> - hybrid_field_type= ((Item_field*) item)->field->type();
>> - else
>> - hybrid_field_type= Item::field_type();
>> + hybrid_field_type= item->field_type();
>>
>> if (check_sum_func(thd, ref))
>> return TRUE;
>
> This is a typical view related bug. So the test cases should be added to
> view.test.
>
Ok, moved to view.test.
> Usually we fix such bugs in the following way:
>
> An occurence of (item->type() [== Item::FIELD_ITEM])
> is substituted for (item->real_item()->type() [== Item::FIELD_ITEM])
>
> So my suggestion is to apply this simple modification:
>
> fix_length_and_dec();
> + item= item->real_item();
> if (item->type() == Item::FIELD_ITEM)
> hybrid_field_type= ((Item_field*) item)->field->type();
> else
> hybrid_field_type= Item::field_type();
>
Ok, done in the new patch.
> I checked that this fix works for your test. Please check that it works for
> test-as3ap as well.
>
Yes, it does.
Best regards,
--
Alexey Kopytov, Software Developer
MySQL AB, www.mysql.com
Are you MySQL certified? www.mysql.com/certification
