Re: bk commit into 5.0 tree (davi:1.2570) BUG#33798

[Davi Arnaut]

Konstantin Osipov wrote:
> * Davi Arnaut <davi@stripped> [08/02/08 14:00]:
> 
>> ChangeSet@stripped, 2008-02-08 08:55:55-02:00, davi@stripped +5 -0
>>   Bug#33798 prepared statements improperly handle large unsigned ints
>>   
>>   The unsignedness of large integer user variables was not being
>>   properly preserved when feeded to prepared statements. This was
>>   happening because the unsigned flags wasn't being updated when
>>   converting the user variable is converted to a parameter.
>>   
>>   The solution is to copy the unsigned flag when converting the
>>   user variable to a parameter and take the unsigned flag into
>>   account when converting the integer to a string.
> 
> The patch is OK to push, see one item below.
> 
>> --- a/sql/item.cc	2008-01-11 15:57:33 -02:00
>> +++ b/sql/item.cc	2008-02-08 08:55:54 -02:00
>> @@ -2580,6 +2580,7 @@ bool Item_param::set_from_user_var(THD *
>>    if (entry && entry->value)
>>    {
>>      item_result_type= entry->type;
>> +    unsigned_flag= entry->unsigned_flag;
> 
> That assignment should be enough -- set() does not truncate
> anything anyway.
>>      if (strict_type && required_result_type != item_result_type)
>>        DBUG_RETURN(1);
>>      switch (item_result_type) {
>> @@ -2875,7 +2876,10 @@ const String *Item_param::query_val_str(
>>  {
>>    switch (state) {
>>    case INT_VALUE:
>> -    str->set(value.integer, &my_charset_bin);
>> +    if (unsigned_flag)
>> +      str->set((ulonglong) value.integer, &my_charset_bin);
>> +    else
>> +      str->set(value.integer, &my_charset_bin);
> 
> In other words, this change does not seem to be necessary.
> 

Yes, it's necessary, otherwise we end-up with a wrong binlog. The
integer to string conversion is different for unsigned and signed, take
a look at String::set.

Regards,

-- 
Davi Arnaut, Software Engineer
MySQL Inc, www.mysql.com

Are you MySQL certified?  www.mysql.com/certification