[Ffmpeg-devel] [PATCH] from DivX, Part 7: MSVC fixes
Thu Jan 26 23:16:21 CET 2006
Michael Niedermayer wrote:
> On Thu, Jan 26, 2006 at 01:04:44AM -0500, Rich Felker wrote:
>> On Thu, Jan 26, 2006 at 12:39:46AM +0100, Michael Niedermayer wrote:
>>> On Wed, Jan 25, 2006 at 01:30:44PM -1000, Steve Lhomme wrote:
>>>> Michael Niedermayer wrote:
>>>>>>> would be cleaner but even then i would say that needs some disscussion
>>>>>>> should be a seperate patch
>>>>>> Given __align8(type, var) was rejected because it's "unreadable", I
>>>>>> don't think this macro would qualify neither. Other than that, I agree
>>>>>> this is a better long-term change.
>>>>> who rejected __align8(type, var)? i didnt or at least i dont remember it
>>>> I think Felker was the one against it.
>> I don't have the authority to reject, just the authority to flame. :)
>>> __align8(type, var) is ok, ill apply it if you send a patch with it
>> I was against it because it's not even obvious that this is a variable
>> declaration without reading the #define.
>> Also, using __ prefix is illegal. __ is reserved for the
>> implementation. For all you know, the C library/compiler might define
>> __align8 for its own purposes. You should use align8 or something not
>> in the reserved namespace at the very least.
> yes fully agree, i missed the __ in there, dunno must have been tired or
> maybe we should borrow the
> #define DECLARE_ALIGNED( type, var, n ) type var __attribute__((aligned(n)))
> from x264?
Here is the patch, I used DECLARE_ALIGNED_8() and DECLARE_ALIGNED_16()
I might be better than having a "random" alignement size, especially
since some sizes are mapped to different values that the one set on some
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
More information about the ffmpeg-devel