On 06/24/2009 11:12 AM, Dave Page wrote:
> On Wed, Jun 24, 2009 at 10:07 AM, Andres Freund<andres@anarazel.de> wrote:
>> On 06/24/2009 10:35 AM, Dave Page wrote:
>>>
>>> On Tue, Jun 23, 2009 at 8:23 PM, Andres Freund<andres@anarazel.de> wrote:
>>>
>>>> I found the issue. The code compares '(unsigned int)(-1)' with
>>>> '(size_t)(-1)' which works on 32bit but not on 64bit.
>>>
>>> Thanks Andres - patch applied.
>>
>> I havent read very much of the code, but at a very quick glance a short
>> policing for issues of this kind looks sensible - unfortunately I cant do
>> this in due time (I think I have spotted some more comparisons of this kind
>> - I am not sure how big the influence of those is though).
>
> Yeah, that certainly wouldn't hurt. If you spot anything in passing,
> please point it out.
I have only looked at the source code of ctlSQLBox - but there is a wild
mixature of unsigned int (32bit, unsigned), int(32bit signed),
long(64bit signed), size_t(64bit signed unsigned) datatypes which are
assigned in most of the possible combinations.
If I have glanced correctly nothing triggers unless somebody gets the
idea to edit a 1GB+ file... But then it probably would trigger on 32bit
as well.
But in my experience if no special care is payed on such issues there
are more bugs hiding...
> We're not going to get a full review done before release now anyway...
Yea, its sounds a bit too invasive to do such a cleanup before release
(it very well could uncover hidden bugs beside causing new ones).
Andres