[FFmpeg-cvslog] r9581 - in trunk/libavcodec: dsputil.c h263.h i386/dsputil_mmx.c mpegvideo_enc.c msmpeg4.h
Ramiro Ribeiro Polla
ramiro
Wed Jul 11 17:34:12 CEST 2007
Aurelien Jacobs wrote:
> On Wed, 11 Jul 2007 11:26:21 -0300
> Ramiro Ribeiro Polla <ramiro at lisha.ufsc.br> wrote:
>
>
>> Aurelien Jacobs wrote:
>>
>>
>>> On Tue, 10 Jul 2007 18:30:37 -0300
>>> Ramiro Ribeiro Polla <ramiro at lisha.ufsc.br> wrote:
>>>
>>>
>>>
>>>
>>>> aurel wrote:
>>>>
>>>>
>>>>
>>>>> Author: aurel
>>>>> Date: Tue Jul 10 22:23:08 2007
>>>>> New Revision: 9581
>>>>>
>>>>> Log:
>>>>> Avoid linking with h263.c functions when the relevant codecs
>>>>> are not compiled in.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>> There are still some missing in libavcodec/dsputil.c, and
>>>> libavcodec/i386/dsputil_mmx.c (and possibly on other arches I haven't
>>>> looked).
>>>>
>>>> ramiro at drake:/usrc/ffmpeg/build$ ../src/configure --disable-muxers
>>>> --disable-parsers --disable-demuxers --disable-encoders
>>>> --disable-decoders --disable-bsfs
>>>> ramiro at drake:/usrc/ffmpeg/build$ make
>>>> [builds everything and gives this error:]
>>>> ramiro at drake:/usrc/ffmpeg/build$ make
>>>>
>>>>
>>>>
>> [error linking due to ff_h263_loop_filter_strength]
>>
>>
>>> Sorry, I can't reproduce this. Could you try a make clean first ?
>>>
>>>
>>>
>>>
>> Reproduced on another machine.
>> Fresh checkout:
>>
>
> Well, I've just tried again with a fresh checkout and the exact same
> configure as you. I still can't reproduce.
> Could you give some details about your compiler, OS...
>
>
Right... I somehow forgot how important that was. I'm running gcc 4.0.3
on an i686 ubuntu (the worst distro to develop on =).
>> Shouldn't h263_v_loop_filter_c and h263_h_loop_filter_c also be under
>> any h263?
>>
>
> No, they shouldn't need to.
> Could you please try the ugly attached patch and tell me if it helps.
>
Yep, that worked. It built fine.
> If it don't, you could also try to add some #if ENABLE_ANY_H263
> around the functions.
> I fear I won't be able to help much as long as I can't reproduce.
>
>
I'll take a closer look later.
Ramiro Polla
More information about the ffmpeg-cvslog
mailing list