00:00soreau: you might be looking for https://wayland.app/protocols/xdg-shell#xdg_toplevel:event:wm_capabilities
00:00soreau: "The compositor will ignore requests it doesn't support."
00:01soreau: nothing indicates that it should be an error
00:08karenw: Excellent. Also I'm getting some weirdness trying to support configure_bounds. When maximizing or fullscreening, kwin is sending a configure with sizes larger than the bounds.
00:09karenw: If the configure is larger than the most recent configure_bounds, I should ignore the bounds?
00:19zamundaaa[m]: karenw: bounds is only for when you decide the size yourself
00:20zamundaaa[m]: It's usually smaller than the fullscreen size
00:22karenw: Fair, I need to track the configure states then. If fullscreen or maximized, I ignore the bounds.
01:22zamundaaa[m]: karenw: no, you should ignore the bounds unless the configure size is zero
01:23zamundaaa[m]: With interactive resize, the user may resize the window to be larger than a single screen
01:23zamundaaa[m]: Preventing them from doing that would be weird
01:24karenw: Hmm... okay. I only choose a size for the initial configure (an arbitary 800x600).
01:24karenw: Aren't resizes normally client-driven in wayland? At least when not using xdg_decoration?
01:33zamundaaa[m]: No
01:34zamundaaa[m]: The client can initiate an interactive resize process, but that's all it does
01:37karenw: Fair enough. I have yet to implement my own CSDs. On mutter/weston et. al. I just set_fullscreen() and pray.
01:45karenw: So other than something like aspect-ratio correction (which I currently don't do, also seems to be impossible when maximized?) I should always honor the exact size sent by the compositor?
01:45karenw: If that size is 0,0, only then do I care about configure_bounds. (which is always the case, before the first time I create a vulkan swapchain)
02:16soreau: regarding aspect-ratio, you can attach a buffer with black borders if you so choose, in the event that the compositor requests a size other than your desired ratio
02:21soreau: for configure event of 0x0, I'd follow what simple-egl does https://gitlab.freedesktop.org/wayland/weston/-/blob/main/clients/simple-egl.c#L610
02:25karenw: Pretty much what I'm now writing. (With a slight niggle that I do separate surface_size and buffer_size for buffer_scale and/or fractional_scale)
02:26karenw: Oh simple-egl does that too, just elsewhere
07:56pq: emersion, Xaver is right. "Colorspace" tells the monitor how to decode (optical) RGB values. It's not specific to HDR, also wide color gamut SDR monitors exists.
07:57JEEB: yea, like if a display was able to tell that it has a display p3 gamut, for example
07:57pq: emersion, in the color-mangement protocol we call it the primary color volume primaries.
08:58karenw: wl_fixes got promoted to core? Neat.
13:48emersion: zamundaaa[m]: thanks for filling all of these MRs for my comments!
14:20emersion: zamundaaa[m]: it seems like kwin sends primaries multiplied by 10'000 instead of 1M
14:21emersion: (or maybe this has changed since the last xx proto?)
14:23zamundaaa[m]: The xx proto has 10k as the multiplier, yes
14:23zamundaaa[m]: My MR to update the protocol has a matching change in the multiplier
14:31emersion: i see, ignore me then!
14:35wlb: wayland-protocols Merge request !379 opened by Vlad Zahorodnii (zzag) members: Add Xaver Hugl as a KWin point-of-contact https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/379 [governance]
14:49emersion: where can i find good defaults for luminance, target primaries/luminance/max_cll/max_fall etc?
14:52mclasen: all this numerology. magic multipliers :)
14:53soreau: I propose that it be renamed to magicpixels protocol
14:55emersion: is the "container" terminology pq-approved?
15:16zamundaaa[m]: emersion: the default luminance levels are defined in the protocol, in the `set_luminances` request and the transfer function descriptions
15:17zamundaaa[m]: It probably wouldn't hurt to add a description to the transfer function enum that specifies them there
15:17emersion: oh, i see!
15:18emersion: yeah, maybe move "The default luminances are" to the transfer_function enum overview?
15:18emersion: since it's useful for other messages as well, and the defaults for st2084_pq are there as well
15:19zamundaaa[m]: About the container terminology, the protocol calls it "container color space" in one place. Might be worth calling it that in set_primaries too
15:20zamundaaa[m]: "Encoding color space" would be another one I find to make sense intuitively. Dunno which one is more technically correct
15:22emersion: makes sense!
15:23Lucretia: why do cursors keep disappearing?
15:24Lucretia: and when calibrating a tablet on kde, it's totally invisble.
15:31karenw: When I touch the screen on KDE, it hides the mouse cursor until I move the pointer again.
15:31karenw: I assume that's what you are seeing?
15:37Lucretia: nope
15:37Lucretia: when using just the mouse, the cursor disappears and refuses to come back
15:38Lucretia: I constantly have to shake it to get it to show and that's after what feels like minutes of trying.
15:38Lucretia: this isn't hte only thing, wayland locks up ff all the time
15:38Lucretia: I've seen a bug for that
15:39Lucretia: and man, does it feel slow, this is a new machine, not a lot spec either.
15:39Lucretia: s/lot/low/
15:39davidre: Please report to us
15:39davidre: as in kde bugs.kde.org
15:40davidre: this is nto an issue with the wayland protocol
15:40Lucretia: davidre: shouldn't you also see a cursor for pen when calibrating?
15:41davidre: I dont know
15:41Lucretia: so you can see where you are clicking, pretty sure it did that on x
15:43Lucretia: and which product does the bug go under?
15:44davidre: kwin, if it's in the wrong place people will move it to the correct one
17:02wlb: weston Issue #990 opened by Derek Foreman (derekf) Follow-up from "Add perfetto instrumentation to weston" with a systemd unit to launch perfetto https://gitlab.freedesktop.org/wayland/weston/-/issues/990 [Debug]
17:25wlb: weston/main: Derek Foreman * 9 commits https://gitlab.freedesktop.org/wayland/weston/compare/0c31f7dfd28a6b5684652cdeed7f1babc47805e2...e22a11513d3743eafedae2bee06dbdd71a735564
17:25wlb: weston Merge request !1670 merged \o/ (Add perfetto instrumentation to weston https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1670)
18:42pq: emersion, for terminology, it would be nice to follow https://gitlab.freedesktop.org/pq/color-and-hdr/-/blob/main/doc/glossary.md
18:42pq: we've tried to do so
18:47pq: zamundaaa[m], ^