[FFmpeg-cvslog] r29009 - trunk/libswscale/swscale.c

Cédric Schieli cschieli
Sat Mar 21 12:12:59 CET 2009


2009/3/21 M?ns Rullg?rd <mans at mansr.com>:
> C?dric Schieli <cschieli at gmail.com> writes:
>
>> 2009/3/21 M?ns Rullg?rd <mans at mansr.com>:
>>> sdrik <subversion at mplayerhq.hu> writes:
>>>
>>>> Author: sdrik
>>>> Date: Fri Mar 20 15:01:51 2009
>>>> New Revision: 29009
>>>>
>>>> Log:
>>>> Use a simpler and more general check for the gray case in the
>>>> planarCopy function
>>>>
>>>> Modified:
>>>> ? ?trunk/libswscale/swscale.c
>>>>
>>>> Modified: trunk/libswscale/swscale.c
>>>> ==============================================================================
>>>> --- trunk/libswscale/swscale.c ? ? ? ?Fri Mar 20 00:33:35 2009 ? ? ? ?(r29008)
>>>> +++ trunk/libswscale/swscale.c ? ? ? ?Fri Mar 20 15:01:51 2009 ? ? ? ?(r29009)
>>>> @@ -1989,11 +1989,8 @@ static int planarCopy(SwsContext *c, uin
>>>> ? ? ? ? ?int y= ? ? ?plane==0 ? srcSliceY: -((-srcSliceY)>>c->chrDstVSubSample);
>>>> ? ? ? ? ?int height= plane==0 ? srcSliceH: -((-srcSliceH)>>c->chrDstVSubSample);
>>>>
>>>> - ? ? ? ?if ((isGray(c->srcFormat) || isGray(c->dstFormat)) && plane>0)
>>>> - ? ? ? ?{
>>>> - ? ? ? ? ? ?if (!isGray(c->dstFormat))
>>>> + ? ? ? ?if (dst[plane] && !src[plane])
>>>> ? ? ? ? ? ? ? ? ?fillPlane(dst[plane], dstStride[plane], length, height, y, 128);
>>>> - ? ? ? ?}
>>>> ? ? ? ? ?else
>>>> ? ? ? ? ?{
>>>> ? ? ? ? ? ? ?if (dstStride[plane]==srcStride[plane] && srcStride[plane] > 0)
>>>>
>>>
>>> This is breaking regression tests.
>>
>> Yes. This was wrong. A proper fix is discussed here :
>> http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2009-March/065915.html
>
> So are you going to fix it?
>
> --
> M?ns Rullg?rd
> mans at mansr.com
> _______________________________________________
> ffmpeg-cvslog mailing list
> ffmpeg-cvslog at mplayerhq.hu
> https://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-cvslog
>

Reimar's commits (r29018 and r29021) should have fixed it by now.

Sorry for the mess.




More information about the ffmpeg-cvslog mailing list