On Sat, Feb 14, 2015 at 11:09 AM, Diego Biurrun <[email protected]> wrote:
> On Sat, Feb 14, 2015 at 05:05:56PM +0100, Luca Barbato wrote:
>> On 14/02/15 16:47, Diego Biurrun wrote:
>> >On Sat, Feb 14, 2015 at 04:40:30PM +0100, Luca Barbato wrote:
>> >>On 14/02/15 16:38, Diego Biurrun wrote:
>> >>>the flags variable as well?  And why not int32_t or ssize_t?
>> >>
>> >>because, as you can read you read a 32bit quantity and then subtract a 
>> >>value
>> >>that is at most 1024.
>> >
>> >That sort of answers one of my questions.  What about flags?  And why
>> >not ssize_t?  I think we should start using more (s)size_t all over
>> >the place.
>>
>> We should not do that mindlessly, if the value returned is an explicit
>> 32bit, is better to use explicit types.
>>
>> (IIRC on certain platforms ssize_t might be less than the 64bit needed to
>> avoid the overflow...)
>
> What overflow?

You could check the CID number, here is the link with the whole flow
https://scan5.coverity.com:8443/reports.htm#v36633/p10274/fileInstanceId=67907580&defectInstanceId=20652447&mergedDefectId=1041122
(you need to be logged).

Not sure about that ssize_t but imho it's a separate thing from this patch.
-- 
Vittorio
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to