[vlc-devel] [PATCH v3] Add draining support to filters

Steve Lhomme robux4 at ycbcr.xyz
Tue Oct 20 14:03:16 CEST 2020


On 2020-10-20 13:53, Rémi Denis-Courmont wrote:
> Ok well that just means that the pull request description is invalid, so 
> -1 either way.

There is no pull request.

More seriously, I disagreed with your previous arguments that the 
draining of the last picture is fundamentally different than the 
draining of "in-flight" or "pending" pictures. And now I have code to 
show that there doesn't need to be 2 code pathes to do almost the same 
thing.

> Le 20 octobre 2020 14:35:01 GMT+03:00, Steve Lhomme <robux4 at ycbcr.xyz> a 
> écrit :
> 
>     On 2020-10-20 13:26, Rémi Denis-Courmont wrote:
> 
>         Based on the description this has the exact same problems as the
>         previous version.
> 
>         -1
> 
> 
>     Not a valid code review.
> 
>         Le 20 octobre 2020 14:04:05 GMT+03:00, Steve Lhomme
>         <robux4 at ycbcr.xyz> a
>         écrit :
> 
>         I'm not sending the 35 patches to the ML (too big) but the
>         individual
>         patches can be found at
>         https://code.videolan.org/robUx4/vlc/-/tree/filters/refactor-ops/13
> 
>         Changes since v2:
>         - rebased on master + epileptic filter support + picture
>         dropping filters
>         - "full" draining of filters is supported in the vout, transcode,
>         mosaic, SDI
>         - a flag to the drain callback of filters is added to tell them they
>         should not only sending the delayed pictures they may have left
>         to send
>         but also anything else based on the source pictures they were
>         given. No
>         more pictures will be given until a flush or a deletion of the
>         filters.
>         - for now no filters give extra pictures, this is left to
>         implement per
>         filter, in addition of the drain callback they already have
>         - the filter chain doesn't store "pending" pictures anymore. When a
>         filter_chain_VideoFilter() is done without (enough) draining
>         before, it
>         starts draining all filters first. This is how the filter chain
>         worked
>         so far. They were stored in pending and if nothing drying the
>         pending
>         chain, they were released in the next filter_chain_VideoFilter()
>         call.
> 
>         Changes since v1:
>         - fix misc compilation issues
>         - rework the fps code so that the extra pictures can actually be
>         added
>         - rework the fps draining with more explanation. Using the
>         p_sys->p_previous_pic->date
>         was incorrect as it was shifted already.
>         - keep the reference date for VDPAU x2 deinterlacing draining
>         - add truncated commit text in last patch
>         ------------------------------------------------------------------------
>         vlc-devel mailing list
>         To unsubscribe or modify your subscription options:
>         https://mailman.videolan.org/listinfo/vlc-devel
> 
> 
>         -- 
>         Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez
>         excuser
>         ma brièveté.
>         ------------------------------------------------------------------------
>         vlc-devel mailing list
>         To unsubscribe or modify your subscription options:
>         https://mailman.videolan.org/listinfo/vlc-devel
> 
>     ------------------------------------------------------------------------
>     vlc-devel mailing list
>     To unsubscribe or modify your subscription options:
>     https://mailman.videolan.org/listinfo/vlc-devel
> 
> 
> -- 
> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser 
> ma brièveté.
> 
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel
> 


More information about the vlc-devel mailing list