[FFmpeg-cvslog] r24260 - in trunk/tests: fate2.mak ref/fate/msmpeg4v1

Måns Rullgård mans
Sat Jul 17 21:39:40 CEST 2010


Michael Niedermayer <michaelni at gmx.at> writes:

> On Fri, Jul 16, 2010 at 11:20:20PM +0100, M?ns Rullg?rd wrote:
>> Michael Niedermayer <michaelni at gmx.at> writes:
>> 
>> > On Fri, Jul 16, 2010 at 03:39:54PM +0100, M?ns Rullg?rd wrote:
>> >> Michael Niedermayer <michaelni at gmx.at> writes:
>> >> 
>> >> > On Fri, Jul 16, 2010 at 11:34:39AM +0100, M?ns Rullg?rd wrote:
>> >> >> vitor <subversion at mplayerhq.hu> writes:
>> >> >> 
>> >> >> > Author: vitor
>> >> >> > Date: Fri Jul 16 12:19:21 2010
>> >> >> > New Revision: 24260
>> >> >> >
>> >> >> > Log:
>> >> >> > Revert r24258:
>> >> >> >
>> >> >> > Log:
>> >> >> > Add msmpeg4v1 regtest
>> >> >> >
>> >> >> > Added:
>> >> >> >    trunk/tests/ref/fate/msmpeg4v1
>> >> >> > Modified:
>> >> >> >    trunk/tests/fate2.mak
>> >> >> >
>> >> >> > According to Mans, "make test" tests already msmpeg4v1.
>> >> >> 
>> >> >> Don't believe anything I say.  It was mislabeled, and is testing v3.
>> >> >> 
>> >> >> What's worse, the encoder or decoder appears to be quite broken, so I
>> >> >> can't add a proper test for it either:
>> >> >
>> >> > i dont want to break peoples dreams and hopes but i do not think we ever
>> >> > had a msmpeg4v1 encoder. I felt it would have been wasted time to write/
>> >> > finish it. Supporting the later v2 and v3 seemed sufficient
>> >> 
>> >> So would you approve a patch to rip out the non-working encoder?
>> >> We should either fix it or remove it.
>> >
>> > iam ok with removing the struct
>> 
>> And any code conditional on msmpeg4v1?  I see no point keeping that either.
>
> that code would be quite helpfull to someone who wanted to finish
> the encoder and someone who wanted to do this likely would not find
> it if its just in a distant past revission

Do you honestly think anyone will ever do that?

> we also could leave the struct and mark it with CODEC_CAP_EXPERIMENTAL

The code is sprinkled with if(s->msmpeg4_version==1) making all the
working encoders slower for no good reason.

-- 
M?ns Rullg?rd
mans at mansr.com



More information about the ffmpeg-cvslog mailing list