[FFmpeg-devel] [PATCH]Fix building with --disable-everything

[Paul B Mahol]

On 2/9/13, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> Hi!
>
> Paul B Mahol <onemda <at> gmail.com> writes:
>
>> > +++ b/libavcodec/x86/Makefile
>> > @@ -62,8 +62,7 @@
>> >                                    x86/h264_weight_10bit.o
>> >  YASM-OBJS-$(CONFIG_H264PRED)   += x86/h264_intrapred.o          \
>> >                                    x86/h264_intrapred_10bit.o
>> > -YASM-OBJS-$(CONFIG_H264QPEL)   += x86/h264_qpel_8bit.o          \
>> > -                                  x86/h264_qpel_10bit.o
>> > +YASM-OBJS-$(CONFIG_H264QPEL)   += x86/h264_qpel_10bit.o
>>
>> Why?
>
> (I am not sure I understand the question.)
> I removed the conditional compilation of h264_qpel_10bit.o
> because it has to be compiled unconditionally if
> compilation with --disable-everything should be supported.

Why? What depends on it?

> (It would not be important atm, but leaving the bug unfixed
> makes future regression tests much more difficult, that is
> why I want to fix it asap.)
>
>> > @@ -82,5 +81,6 @@
>> >  YASM-OBJS                      += x86/dsputil.o                 \
>> >                                    x86/deinterlace.o             \
>> >                                    x86/fmtconvert.o              \
>> > +                                  x86/h264_qpel_8bit.o          \
>>
>> This looks wrong.
>
> Is there another way to force unconditional compilation?
>
> Thank you, Carl Eugen
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>