GNOME 42 Released With Many Improvements From Wayland To GTK4 Porting

[dimko]

By applying pressure to Wayland community.
Wayland has been developed for over a decade and has nothing to show for it. Yeah it may work and yeah it may have more secure GUI over RDP, but i dont use guy for RDP. i am happy with SSH. i want my gui to take advantage of hardware I have. We Linux users are always 2-3 steps behind, its annoying. Gnome could implement some GUI calls, something ahead of wayland in terms of HDR. Windows and Mac already have it, you can guess based on their API on how it will work anyway. So when HDR becomes available in wayland some skeleton for HDR is there in Gnome.

[RahulSundaram]

By applying pressure to Wayland community..

This does absolutely nothing. There are a couple of ways to actually influence development a) Become contributors - bug reports are a good starting point b) Funding - if individual developers are raising funds or an organization is behind it, you could donate or if there is a commercial product that is related, buy that product.

[Spacefish]

How do you propose they add this when Wayland doesn't support HDR yet and Xorg doesn't support it and probably never will?

It´s actively worked on in wayland.. IMHO gnome is not required to support HDR, wayland needs to support it, once that´s done, the application can notify wayland if it want´s it surfaces to be 10bit / apply a HDR Gamma Curve on them..
As far as i understand the current development in wayland, they want to combine Color Management and HDR and release them together, somehow makes sense, as gamma curves and translation between two colorspaces is almost the same operation, you get an input 3 channel value and generate an ouput 3 channel value via a LUT or some function.

[lumks]

It's funny how the release notes in Phoronix seem to be infinitely more interesting than the official ones.
Is it a symptoms of the priorities of the developers of Gnome being completely misaligned with my own desires?

Yes. The official ones are made for the normal non-techy user that gnome targets. So not much infrastructural talk to find there and that while 42 is mostly a infrastructural release with only a cherry on the top to see for normal users.

[sebastianlacuesta]

Using on Debian and looks nice. Except that thanks to my nvidia gpu wanting an xserver and not having the option to choose between wayland and x11 in gdm anymore. Yes, now I can use optimus again (bumblebee broke long time ago and needed to remove it), but lost the nice multitouch gestures because of the x11 dinosaur. Hope some day nvidia play nice with XWayland and X11 would be a thing of the past for most software.

[You-]

Is it a symptoms of the priorities of the developers of Gnome being completely misaligned with my own desires?

It is a symptom of the release notes being collected long after the features have been developed along with a few developers working at full steam and not having a chance to breathe and type their work up. Some sort of tagging or other system may be useful to store the relevant changes.

[You-]

And still no HDR...

MVP is planned for Gnome 43. Some of the work is already in place and EG high depth pngs recently got tested out due to gtk having the code paths in place (and it exposing a bug that was quickly squashed).

I bought a HDR monitor in anticipation of sampling the work from the next release cycle

[Myownfriend]

.
As far as i understand the current development in wayland, they want to combine Color Management and HDR and release them together, somehow makes sense, as gamma curves and translation between two colorspaces is almost the same operation, you get an input 3 channel value and generate an ouput 3 channel value via a LUT or some function.

Yup, you're correct. I actually love the idea that they're working on them in tandem. As far as I'm aware though, the finalization of the spec in Wayland is still a way off and but they're actively experimenting with implementation in Weston as they usually do. There's also an MR for Mutter that was made in anticipation of the spec being finalized though.

[ⲣⲂaggins]

Well that video was positively epilepsy-inducing...

• Input handling has been significantly enhanced, resulting in lower input latency and improved responsiveness when the system is under load. This will be particularly beneficial for games and graphics applications.

That's great, but what they didn't mention is that if your program is unfortunate enough to be using GTK, then as of GTK4, you are forced to opt into event compression - meaning you never actually see the low-latency benefits of this; the toolkit greedily keeps them all for itself.

Event compression is always enabled in GTK 4, for both motion and scroll events. If you need to see the uncoalesced motion or scroll history, use gdk_event_get_history() on the latest event.

[ⲣⲂaggins]

Have the text rendering regressions in GTK4 been fixed yet?
I'll check it myself in the coming days. There hasn't been any activity lately on the relevant threads at gitlab.gnome.org/GNOME/gtk
The two regressions are
1.) Subpixel rendering has been dropped
2.) Subpixel positioning is on by default
Regression Home discards nearly 2/3 of the horizontal resolution on standard LCD displays (the overwhelming majority of displays on the market today).
Regression Forum discards roughly half the horizontal and vertical resolution of all displays.
There are valid situations where it's useful to do Home or #2, but text for reading isn't one of them. Even though most people think they're immune to eye strain, studies have shown that almost all people suffer from eye strain if they spend long periods (ie hours per day) at a computer.

Ah, another Baggins.

It seems pretty clear from the bug reports and MR that the developers consider the text rendering regressions a feature, not a bug, and so they probably won't ever be fixed. This came up in the last thread about it:

https://www.phoronix.com/forums/forum/software/desktop-linux/1310381-gnome-42-beta-released-begins-the-ui-feature-api-freeze-more-apps-ported-to-gtk4/page4