[FFmpeg-devel] [PATCH] Fix non-rounding up to next 16-bit aligned bug in IFF decoder

Sebastian Vater cdgs.basty
Mon May 3 15:17:38 CEST 2010


Sebastian Vater a ?crit :
> Benoit Fouet a ?crit :
>   
>> On Wed, 28 Apr 2010 23:24:55 +0200 Sebastian Vater wrote:
>>   
>>     
>>> Reimar D?ffinger a ?crit :
>>>     
>>>       
>>>> On Wed, Apr 28, 2010 at 03:51:26PM +0200, Sebastian Vater wrote:
>>>>   
>>>>       
>>>>         
>>>>>> This is not how people commonly round up to 2. Most people do
>>>>>> (x+1)&~1, so you could do:
>>>>>>       
>>>>>>           
>>>>>>             
>>>> Note I think this does not apply here, but you definitely should _not_
>>>> use (x+1)&~1 in FFmpeg code, use FFALIGN instead.
>>>>   
>>>>       
>>>>         
>>> Fixed. See attached patch!
>>>
>>>     
>>>       
>> as Reimar said, it probably does not apply here.
>> The usage for FFALIGN is, e.g.:
>> foo = FFALIGN(foo, 16);
>>
>> to align foo on a 16 boundary.
>>   
>>     
>
> Yes, this indeed looks better...see new attached patch, that should do
> it the best way.
>   

Oops, I acidentally did a delete of the comment from the pixel width
alignment line, sorry.

So here, again, Fixed!

-- 

Best regards,
                   :-) Basty/CDGS (-:

-------------- next part --------------
A non-text attachment was scrubbed...
Name: iff-decoder-fix.patch
Type: text/x-diff
Size: 1212 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20100503/46dc9858/attachment.patch>



More information about the ffmpeg-devel mailing list