[vlc-devel] [PATCH 1/4] mmal: update the code from the one actually used in Raspbian

[Steve Lhomme]

On 2020-01-18 8:40, Rémi Denis-Courmont wrote:
> Le lauantaina 18. tammikuuta 2020, 7.18.54 EET Steve Lhomme a écrit :
>> This is a hack to know which display id is used when going in full-screen.
> 
> It won't work anyway. There's no guarantees that the Qt plugin is providing
> the window, or if it is, how it counts displays.

In this current iteration of the code you're correct. I'll just remove 
that part of the code.

The original code has a trick (fake display) that forces using the MMAL 
vout rather than X11 or OpenGL when going in fullscreen, for performance 
reasons. That's where the Qt value comes handy so it matches the display 
it's given. But this patchset doesn't include this "feature" so it's not 
needed.

>> It's either HDMI0 or HDMI1 and the display module needs to know that to go
>> in the right mode.
> 
> You're assuming a specific hardware version and Qt back-end there. I don't know
> if Raspian patched Qt to guarantee that Qt provides the same numbers as XRR
> (on local display). But it's not generally valid. E.g., I've been using Ubuntu
> on rPI3 since Raspian can't be bothered to support 64-bits, and they AFAIK,
> they just ship normal Qt, not some rPI-specific fork of it.
> 
>> That corresponds to the Qt variable that can use a
>> different display when going in full-screen. I suppose it takes the widget
>> with the VLC window and send it to the other screen. But the display module
>> isn't told about it.
> 
> The display module can find what subrectangle of the window is visible on each
> output by monitoring the window and output events. It's just not "supposed" to
> care. E.g. what will it do if two active outputs are clones?
> 
> -- 
> 雷米‧德尼-库尔蒙
> http://www.remlab.net/
> 
> 
> 
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel
>