ChanServ changed the topic of #wayland to: https://wayland.freedesktop.org | Discussion about the Wayland protocol and its implementations, plus libinput
yin has quit []
yin has joined #wayland
AJ_Z0 has quit [Read error: Connection reset by peer]
coldfeet has quit [Remote host closed the connection]
ammen99 has joined #wayland
<ammen99>
a question regarding focus and xdg_popups. Say an app opens a right-click context menu (xdg_popup) which has an active grab. Should the wl_keyboard enter the popup, or stay at the main surface? I am specifically debugging one issue in Wayfire (https://github.com/WayfireWM/wayfire/issues/2600), where this seems to cause issues and I am wondering whether it is a bug in Wayfire or in Qt.
MrCooper__ is now known as MrCooper
kts has joined #wayland
<jadahl>
ammen99: "while the top most grabbing popup will always have keyboard focus" <-- this is in the spec
<ammen99>
jadahl, thanks, so this sounds like a Qt bug.
fmuellner_ has quit [Remote host closed the connection]
Solomoncyj has joined #wayland
feaneron has joined #wayland
<Solomoncyj>
what differcece does a pip window have vs a normal window other then always on top?
fmuellner has joined #wayland
<SardemFF7>
depending on the DE, you could want the pip window to switch workspace/desktop/whatever concept you have for a group of windows, or not
<MrCooper>
and be visible at all times, even e.g. during Alt-Tab / gnome-shell's overview / ...
<Solomoncyj>
the implemntation is up to the compositor right? eg i want pip to show up as a normal window in alt-tab
novenary has quit [Remote host closed the connection]
novenary has joined #wayland
<SardemFF7>
it could even be special-cased to some hardware overlay that send the surface to a specific second monitor input and that’s the monitor doing the actual PIP
<kennylevinsen>
solomoncyj: it will be as defined in the protocol when it is finalized. In general, the compositor decides many things and has the last say as it represents the user and their preference.
<kennylevinsen>
"alt-tab" is not part of any protocol, but is a compositor feature that the compositor decides how works. sway has no such thing for example.
<MrCooper>
the general point is that "this is a PiP window" gives the compositor more information than "this window should be always on top", so it can treat it more like the user expects
<kennylevinsen>
There *could* be other differences to a normal window. E.g., some platforms restricts what a pip window can do, what UX it should have and what controls are possible.
<kennylevinsen>
Follow the protocol discussions for more info :P
<emersion>
good ol' "policy, not mechanism"
narodnik has quit [Ping timeout: 480 seconds]
narodnik2 has quit [Ping timeout: 480 seconds]
<Solomoncyj>
also can an appliaction self-declare the keys it wants for global shortcuts?
fmuellner has left #wayland [#wayland]
<Solomoncyj>
what is adg-shell for?
narodnik has joined #wayland
fmuellner has joined #wayland
<jadahl>
solomoncyj: in the global shortcuts portal, applications can hint about preferred keyboard combinations, but in the end, it's expected to be controlled by the user
narodnik2 has joined #wayland
<Solomoncyj>
so there is no ther way but for the user to go to the settings app to change it?
narodnik has quit []
narodnik has joined #wayland
<emersion>
pq: how does `struct di_hdr_static_metadata` map to the protocol? in particular, how does desired_content_max_luminance map to target_luminance.max_lum and target_max_cll? are all these the same?
<emersion>
or is the EDID missing one of these values?
<pq>
hmm
<pq>
emersion, I think the display "desires" should be the target color volume for compositors and clients.
<emersion>
hm, right
<emersion>
i guess i can rephrase my question then :P
<pq>
so yeah, I'd do like you said.
<emersion>
what is the difference between target_luminance.max_lum and target_max_cll?
<pq>
ah
<emersion>
tbqh, i'm not sure i understand target_max_cll
<selckin>
a/21
<pq>
HDR static metadata does not have a target_max_lum, it's the maxCLL instead.
<pq>
While target_max_lum is about the MDCV (what the mastering display was capable of), maxCLL is measured from the material and can be recovered by inspecting all pixels.
<pq>
They are very similar concepts, enough that if one is missing and the other is needed, I might substitute one for the other.
<pq>
With mastering maximum luminance the difference between MDCV and measured pixels is probably not meaningful. We should still have the reference white level for putting everything into "perspective".
<pq>
Perhaps the difference is easier to understand from a different example.
<pq>
Let's consider the MDCV gamut (the primaries & white point) vs. the measured pixel statistics gamut (like a bounding box for the actually used colors).
<pq>
Let's say we have a programme that is intentionally authored to be very low saturation, for artistic purposes. This means the pixel bounding box gamut is very small.
<pq>
At the same time, the MDCV gamut can be large, it's just not fully used.
<emersion>
ah, that makes sense
<emersion>
so the container color volume and target color volume are picked up-front
<pq>
yes
<emersion>
while the max_cll and max_fall are computed, additional metadata which can help with tone mapping
<pq>
indeed
<emersion>
thanks!
<emersion>
what does MDCV stands for btw? i grasp what it is, but the letters i'm not sure
<pq>
If we were to add a little "color vibrancy" in the display, we could do it based on either gamut. Doing it based on the pixel gamut would blow out the saturation and destroy the artistic effect. Doing it based on the MDCV would be better.
<JEEB>
emersion: mastering display color volume
<emersion>
oh
<pq>
that ^
<JEEB>
you can interpret it as "mastering person's display did not show anything beyond these parameters"
akyoto has quit [Remote host closed the connection]
<emersion>
ty :)
<JEEB>
for a very long time I was very confused as to why the author's display info was there
<JEEB>
but it's a weirdly indirect gamut/tone mapping hint
akyoto has joined #wayland
<emersion>
is there a reason why we don't just compute this value like max_cll and max_fall?
<pq>
Correct, and when we talk about target MDCV, it means "the viewer's display cannot show anything beyond these parameters".
<emersion>
the primaries, i mean
<pq>
emersion, the reason is what I just explained with the two gamuts example.
<emersion>
aw
<emersion>
i read it the other way around
<emersion>
hm, i see
<pq>
It's a bit contrived example.
<emersion>
with MDCV you see how much intentionally unused color volume space was left by the author
<pq>
If the source's pixel gamut is small, we can simply display it as-is.
<pq>
yes
<pq>
If the material is heavily tinted, that might be a better example. One wants to maintain the tint while squeezing the gamut into the display.
<emersion>
indeed
<pq>
OTOH, could we not do that just by taking into account the white point of the container? I dunno, maybe.
<pq>
hmm, probably not
narodnik2 has quit [Ping timeout: 480 seconds]
narodnik has quit [Ping timeout: 480 seconds]
narodnik has joined #wayland
narodnik2 has joined #wayland
<Solomoncyj>
what does adg-shell do?
<davidre>
I dont know adg-shell, is it a custom protocol of someone?
<pq>
In the above, I implemented parametric<->ICC by doing the simple linkage manually, and letting LittleCMS come up with the actual rendering-intended color transformation between two ICC profiles and then glueing the linkage math on top.
<pq>
IOW, there is no limitation in what kind of ICC profiles can be combined with parametric profiles.
<pq>
...as long as they are Display class.
nerdopolis has joined #wayland
coldfeet has quit [Quit: Lost terminal]
Moprius has joined #wayland
rasterman has quit [Quit: Gettin' stinky!]
gusnan has joined #wayland
feaneron has joined #wayland
garnacho has quit [Quit: garnacho]
garnacho has joined #wayland
<Solomoncyj>
do compositord auto apply "preffred keybinds" for wayland apps that use them?
Solomoncyj has quit [Read error: Connection reset by peer]
fmuellner has quit []
fmuellner has joined #wayland
mripard has quit [Quit: WeeChat 4.6.3]
<kennylevinsen>
what are "preferred keybinds"?
<kennylevinsen>
ah they left
iomari891 has joined #wayland
<MrCooper>
my guess is it's about the global shortcuts portal