[MPlayer-dev-eng] Question regarding persistence in vf_overlay

[Reimar Döffinger]

On Fri, Aug 28, 2009 at 08:33:20AM -0400, Jason Tackaberry wrote:
> On Thu, 2009-08-27 at 16:24 +1000, Aleks Huson wrote:
> > This is all most likely down to my lack of understanding about how this all
> > fits together... but I'd appreciate any light people can shed on this so I
> > know where to look next.
> 
> It's probably more likely this functionality got broken during the most
> recent round of reviews.

If I had to guess, maybe after startup is should treat the shared memory
contents as dirty/changed regardless of the state of the byte used
to update on changes?
Note that this means that an application will have to fill the shared
memory region with valid data before starting MPlayer (previously
I guess it only had to be valid after telling the filter that new
data is available).
And I didn't look at the code so I might of course be completely wrong.