SDL Developers Weigh Reverting Wayland Over X11 For SDL 3.0

[DMJC]

I don't mind which way this goes, as long as SDL3 is better than SDL2 and it's easy to migrate, I'll update my code to remain relevant. I just wish that the desktop/Widget APIs on Wayland worked as nicely as they do on X11. Specifically GTK and tray icons/tray icon menus. At the moment it's a huge step backwards trying to run MATE and other desktops on GTK4/Wayland. I'd like GTK to either add the missing components back on, or have a formal extension system so other desktop groups can extend the API and keep their desktop environments working. There's far too much politicisation for/against Wayland adoption.

[anda_skoa]

Specifically GTK and tray icons/tray icon menus.

Aren't tray icons/menus already handled via D-Bus and thus independent of whether the desktop session is on X11 or Wayland?

Since like ages ago?

[Britoid]

Any alternative to Wayland would end up having the same problems so I'm not quite sure what people are dragging on about.

Is it bad that Wayland development has hurt its real world usage? Yes.
Is it the fault of the protocol itself, no.

This is 100% a people caused problem, Wayland protocol proposals are filled with drama and arguments, including people telling Valve developers their needs are wrong (cause you know, Valve obviously has no idea how games should work).

Wayland is needed because it aligns to how graphics hardware works that allows "modern features", the stubbornness over protocol additions have caused this.

[Volta]

According to https://en.m.wikipedia.org/wiki/X_Window_System the X11 protocol was finalised 1987, three years after initial release.
Wayland had its initial release 15 years ago, and its ecosystem is still not feature complete, comparatively.

How does X11 from 1987 compares to current X11? It's not even remotely comparable. How many workarounds for broken or outdated designs are there? X11 was in development for more than thirty years and it's still not complete. It won't ever be.

[mdedetrich]

HDR wasn't even possible from a technological point of view until very recently. LED backlight, LCD panel and OLED technology were not able to handle the brightness levels required. A proper HDR display needs to achieve peak brightness levels 3-4x higher than a standard "desktop" display.

OLED even with low brightness settings (i.e. 200 nits or above) can easily achieve HDR as the main point of HDR is the contrast and as you know OLED's have perfect blacks. This is also why VESA has different way of rating OLED panels, its just that the tone mapping algorithm is different.

The real problem is that historically OLED has been insanely expensive until the past 5 or so years.

[mdedetrich]

how on earth have i been using wayland these last few years if it is unuseable?

Its unusable for use cases that aren't yours, and desktop has many usecases

[dpeterc]

Some of the features of X11 was not well thought out. These features need to go and the applications that use them will not work in the future without some modifications. This is the way it has to be.

I develop X11 based application which defines its own icon based on its contents.
When users run several instances of my application, they can quickly recognize which instance they want to switch to, by looking at the icon on system tray.
On Wayland, I can't even set my own application icon, much less change it on the fly.
This is possible on all other desktop GUI systems, from Windows to Mac.

Now please explain to me how all the other systems were designed by idiots, and Wayland has been well thought out.

[ssokolow]

typical of computing.
Big Iron from the 60s were replaced by cheaper minis that lacked many features.
minis were replaced by commodity servers, that lacked many features
everything was replaced by x86 which lacked many features
...
and today people are jubilant about virtualisation and 'containers' - which is 60 year old tech....

To be fair, those first four were all about making computing more affordable for a wider range of customers. It's not their fault that it took so long for certain mainframe techs to trickle back into the low-cost tier that people forgot about them.

[ssokolow]

Aren't tray icons/menus already handled via D-Bus and thus independent of whether the desktop session is on X11 or Wayland?

Since like ages ago?

https://freedesktop.org/wiki/Specifications/StatusNotifierItem/

Yeah. KDE started using that spec with KDE 4 because XEmbed doesn't work reliably with things like alpha-blended icon backdrops, displaying icons in an overflow menu, displaying the same icon at multiple sizes in more than one place, etc.

[pracedru]

I develop X11 based application which defines its own icon based on its contents.
When users run several instances of my application, they can quickly recognize which instance they want to switch to, by looking at the icon on system tray.
On Wayland, I can't even set my own application icon, much less change it on the fly.
This is possible on all other desktop GUI systems, from Windows to Mac.

Now please explain to me how all the other systems were designed by idiots, and Wayland has been well thought out.

I didn't say that. And I agree that this is a mistake. The features I wrote about did not include this one.