Regarding the outputdevice Wayland protocol

[Sebastian Kügler]

Hi Drew,

Thanks for reaching out!

On zaterdag 4 november 2017 15:10:33 CET Drew DeVault wrote:
> As part of our effort to reuse existing Wayland protocols for common
> desktop features in Sway (and wlroots), we're taking a look at KDE's
> output-management.xml. However, we feel like the related
> outputdevice.xml protocol has some overlap with wl_output - can you
> provide some context for these protocols and would you consider
> working with us on a simplified protocol we can share?

outputdevice.xml is indeed an extended version of wl_output. It has
some additional options that we need for the outputmanagement stuff, so
it becomes useful, wl_output wasn't sufficient there. I've added the
current mode, edid and enabled fields, as wl_outputs are always
supposed to be enabled, and outputdevices are just connected displays
which may be enabled or not. We also needed a way to pass through the
edid blob. We've checked with upstream wayland, and there was "little
desire" to have these things in wl_output, so an extended version is
what we did. I'm not overly happy with this, but at some point we had
to move on, even if it left warts on the protocol semantics.

A simplified protocol to be shared could be two things:

* merged wl_output with edid, currentmode and enablement added
* simplified outputdevice which uses a wl_output and attaches just the
  above properties

I'm not sure if the above *could* work, since a wl_output was never
designed to be not enabled, so just sitting there, being available, but
the above is at least what I can imagine recalling the discussions we
had back then when we implemented it in kwayland. Martin can probably
provide some more insight here.

Cheers,
-- 
sebas

http://www.kde.org | http://vizZzion.org