[FFmpeg-devel] [PATCH] doc/filters: document the unstability of the shorthand options notation.

Michael Niedermayer michael at niedermayer.cc
Tue Aug 8 01:01:56 EEST 2017

On Mon, Aug 07, 2017 at 11:03:51AM +0200, Nicolas George wrote:
> > >Lets take a step back and look at this
> > >
> > >There are some rarely used options in multi input filters like
> > >overlay which break.
> > >Noone even noticed except me
> > >
> > >And you propose to declare the most used syntax from every filter
> > >unstable.
> > >
> > >This just doesnt add up, its like shooting the patient in the head as
> > >a treatment for a cold
> No, that is like telling the patient they need to take their antibiotics
> until the last day of treatment and not stop once they feel better,
> because they risk getting a pneumonia. Which is technically true, only
> very unlikely.
> Le decadi 20 thermidor, an CCXXV, Marton Balint a écrit :
> > Do you like this text better?
> > 
> > Future evolutions of filters may require inserting new options or changing
> > their order, especially for the non-essential options, and while we make
> > reasonable effort to keep the order so the options given without their name
> > would not break, sometimes that is not feasable. Therefore users should
> > generally favor the @var{key=value} notation, or refer to the Changelog file
> > which should contain such incompatible changes.
> This is obviously the intent, but I do not think that the documentation
> is the place to make that kind of intent statement.
> > Yeah, what is needed for compatibility? Only a single AVOption line, or
> > additional code as well?
> Additional code as well. Not a lot, but still some. And testing. An
> exponential number of cases.
> I do not intend to do it.
> We have already wasted more time than all the users combined will spend
> typing "x=", "y=". Seriously, people !?!
> Now, to all that stated a negative opinion about this:
> I have in the queue a patch series that changes the options in a minor
> way, but at the same time fixes bugs and long-standing limitations of
> lavfi and makes the code more robust and cleaner.
> Do you :

> 1. propose to implement and test compatibility options yourself?

If you push the lavfi patchset but the doc option instability patch
is not pushed then ill likely try to fix the compatibility issue
due to the moved options in overlay and similar filters.

If theres some preferrance on how to fix it?
(iam happy to fix it in the way you perfer as long as its not more
 work than the obvious way to fix it)


Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Why not whip the teacher when the pupil misbehaves? -- Diogenes of Sinope
-------------- 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/20170808/5046498c/attachment.sig>

More information about the ffmpeg-devel mailing list