Frog-FIFO-V1 Aims To Address Mesa's "Fundamentally Broken" Wayland Code

[OneTimeShot]

Isn't this conflating two different things? A FIFO stack bug in Mesa and Wayland protocol development process?

Refactor FIFO, then get the Wayland stuff needed finished and approved. I can't imagine that 15 years into Wayland's existence there's anything that dramatic still to do...

[xpander]

Not only that but there are more things to fix

-Global hotkey support
-Copy-paste between wayland and xwayland that doesn't break
-Drag and drop between applications is finnicky and not always working
-Window management with positions, geometry etc are lacking features.

Just to name some. Hopefully things get better faster. its been 15 years in development.
Lots of people use wayland just fine already. It depends on the needs i guess.

but for me its still a broken mess tbh.

[Old Grouch]

I believe part of the reason upstream Wayland is so slow to add new protocols/features is to try to prevent a 'hot mess' of additional functionality that breaks the design rules/consistency and generates a rapidly expanding technical debt. It is the old 'purity versus pragmatism' tension, and Wayland has tended towards the 'purity' end of the spectrum.

Many people who know what they are doing support this slow development: and many people who also know what they are doing think it is the wrong thing to do. The end result is a great deal of loud statements of a position, and a great deal of frustration.

There are a number of things that are easy to do in the X Window System which are difficult or (possibly by design) impossible to do using Wayland, which means there will be a group of people for whom Wayland does not meet their use-cases. How such people's needs will be met in future is an open question.

[WileEPyote]

That's a pretty damning callout. I love it.

My message to Wayland:
Nelson_Ha-Ha.webp

[bug77]

Helping Linux desktop a lot. That's for sure.

Well, the bar isn't that high... But yes, the difference between having to get things done and waiting for planets to align.

[pokeballs]

> Swapchain starvation caused by being required to stall for the 'frame' event. This leads to it only ever being possible to have 1 image in flight at a time, which starves the GPU of work and can lead to lower performance and stutters.

You can call that "huge gaping issues" all you want, but the reality is that X never really supported multiple displays (no VRR, no mixing frame rates...).
So Wayland is already a much better alternative in the world we currently live in.

[Britoid]

> Swapchain starvation caused by being required to stall for the 'frame' event. This leads to it only ever being possible to have 1 image in flight at a time, which starves the GPU of work and can lead to lower performance and stutters.

You can call that "huge gaping issues" all you want, but the reality is that X never really supported multiple displays (no VRR, no mixing frame rates...).
So Wayland is already a much better alternative in the world we currently live in.

X being a mess is no excuse to sabotage people trying to use it's replacement.

The Steam Deck is arguably the most popular commercial PC that ships with Wayland right now.

[Estranged1906]

IBM-GNOME losers are SEETHING about this lmao

[PennRobotics]

part of the reason upstream Wayland is so slow to add new protocols/features is to try to prevent a 'hot mess' of additional functionality

To recap: X sucks ass on accident. Wayland sucks ass on purpose.

[royce]

Wlroots has been doing this for ages, they have their own batch of wlr-* protocols for use today on sway (and often picked up by other compositors) while they're getting upstreamed into wayland proper:

https://gitlab.freedesktop.org/wlroots/wlr-protocols/-/tree/master/unstable?ref_type=heads

The eagle-eyed will recognise many of these as already upstreamed into wayland.