[FFmpeg-devel] AVClass & AVOption [VOTE]

Michael Niedermayer michael at niedermayer.cc
Mon Jun 6 23:07:22 CEST 2016


On Sun, May 29, 2016 at 01:32:54AM +0200, Michael Niedermayer wrote:
> Hi
> 
> It was suggested in the IRC meeting today that i start a vote to
> resolve if AVClass & AVOption should be added to AVCodecParameters
> This question needs to be awnsered before the next release because
> the ABI would be broken if its added afterwards
> the lack of any decission blocks the release which is worse than
> either decission, otherwise a vote might be silly for such technical
> question but eve a bad decission is better than no decission here
> 
> 
> The disadvanatges are:
> 1 more field in the struct
> a high level API that some feel is unneeded for AVCodecParameters
> it could be confusing to some that there are 2 different ways to
> access the fields
> people might use it as av_log() context when they intended to use a
> different context for it
> There are probably more please help fill the list if you know more
> 
> The advanatges are:
> More consistent availability of AVOptions and AVClass in public
> structs
> Makes supporting multiple FFmpeg versions and distros easier and
> cleaner for applications. (reduces lists of #if version checks)
> Provides default/min/max values, allows basic validity checking within
> min/max
> Avoids mysterious crashes if an application uses avoption functions
> on AVCodecParameters
> Introspection
> Serialization support that may be usefull for ffserver or an application
> replacing ffserver
> 
> 
> An application that doesnt want to use AVOptions or AVClass can
> completey ignore them.
> 
> Please state clearly if you agree to add AVClass&AVOption to
> AVCodecParameters or if you disagree about adding it or if you dont
> care either way
> 
> The vote will end 1 week from now, simple 50% majority (Yes vs no)
> I dont really know who should be eligible for voting, so i suggest
> everyone from the vote comittee but iam happy with anything, just
> stating somethng ...

It seems the vote is a tie 4:4
BBB suggested to add another week, i agree, vote extended by 7 days
that shouldnt affect the release as more time is needed to fix
regression bugs anyway

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Into a blind darkness they enter who follow after the Ignorance,
they as if into a greater darkness enter who devote themselves
to the Knowledge alone. -- Isha Upanishad
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20160606/bf70ee46/attachment.sig>


More information about the ffmpeg-devel mailing list