<div dir="ltr"><br>
<br>
2017-06-28 22:10 GMT+02:00 Rémi Denis-Courmont <<a href="mailto:remi@remlab.net" target="_blank">remi@remlab.net</a>>:<br>
> Le keskiviikkona 28. kesäkuuta 2017, 20.06.31 EEST Mathieu Velten a écrit :<br>
>> Le mer. 28 juin 2017 à 21:47, Rémi Denis-Courmont <<a href="mailto:remi@remlab.net" target="_blank">remi@remlab.net</a>> a<br>
>><br>
>> écrit :<br>
>> > Le keskiviikkona 28. kesäkuuta 2017, 21.22.06 EEST Mathieu Velten a écrit<br>
> :<br>
>> > > The code to initiate the VADisplay has been moved in hw/vaapi.<br>
>> > > You can use --vaapi-display-type to force a type of display.<br>
>> ><br>
>> > The display type is normally chosen by the vout window provider; the<br>
>> > semantics<br>
>> > of that option do not seem to make much sense.<br>
>> ><br>
>> > Also the VA-DRM module shall not depend on X11. This patch is a<br>
>> > regression.<br>
>> ><br>
>> > Nack.<br>
>><br>
>> Then what do you think about:<br>
>> - removing the code regarding VADisplay initialisation from the init of the<br>
>> gl vaapi converter<br>
>> - fetch it later just before the first conversion through the existing<br>
>> instance singleton<br>
><br>
> Your design is already broken to start with if you assume the VA instance is a<br>
> singleton.<br>
<br>
Is it not the current design ?<div>Currently for DR the VADisplay is shared between the vout and the decoder through something looking a lot like a singleton (instance.c). I am bad at C so I could be wrong.<br>
<br>
><br>
> --<br>
> 雷米‧德尼-库尔蒙<br>
> <a href="https://www.remlab.net/" rel="noreferrer" target="_blank">https://www.remlab.net/</a><br>
><br>
> _______________________________________________<br>
> vlc-devel mailing list<br>
> To unsubscribe or modify your subscription options:<br>
> <a href="https://mailman.videolan.org/listinfo/vlc-devel" rel="noreferrer" target="_blank">https://mailman.videolan.org/listinfo/vlc-devel</a><br>
</div></div>