I don't know the implementation details but I can't really complain about the state of wayland today. It used to be annoying to get working many years ago (worse because I had a nvidia gpu). But today I drive a nigthly build of niri, run it by just spawning an dbuss session and everything works. Bluetooth audio, screen sharing, fractional scaling, no tearing, no font blurring. Every utility I needed has been created and works quite nicely (e.g.: wdisplay). I can even play video games with HDR support.
I have a more stable experience with wayland today than I had with x11. Which to be fair was not only because of wayland but because desktop linux as a whole has made a lot of progress in the last years
I don't think it's true that anything is architecturally or fundamentally broken in Wayland (though if you disagree, I'm very curious what you think is so deeply broken).
Most of the issues and slow adoption were because the core protocol was deliberately kept extremely minimal, and agreeing on all the needed extensions took a long time. Don't take it from me, but rather from KDE developer Nate Graham: https://pointieststick.com/2023/09/17/so-lets-talk-about-thi...
As such, anyone who tried it early probably had to deal with a pretty large amount of non-working stuff, but by now the platform is capable of most features people require and the biggest remaining bottleneck is that software needs to use these new APIs.
Window positioning is one that on its own is sufficient to make me ignore Wayland, as it means that without my own compositor with my own extension, I can't get a file manager that will behave how I want it.
Most people won't care, but for a number of us Wayland is stubbornly refusing to support functionality we see as dealbreakers.
That's fair! I believe that window positioning also works on XWayland, though, so running your file manager that way should still work with the rest of the system being Wayland (and Gnome has no plans to drop XWayland afaik).
I believe the main holdup is a desire for Wayland to be usable with e.g. VR interfaces where there is no simple 2d grid.
Out of curiosity, how do you want the file manager to behave? And did you write your own or are you using an existing one that works that way?
It's managing the desktop too, so I'm not sure that'd work unless running Xwayland in "rootful" mode, in which case I might just as well run X.
The VR stuff is a poor excuse - just fail on that scenario. Nobody that cares about window positioning will have an issue with that.
My file manager defaults to re-opening a window for any directory to a previously snapshotted location, like the Amiga Workbench did. And, yes, I wrote my own. It's a few hundred lines of of a quick and dirty Ruby hack talking directly to a pure Ruby X11 binding, which is anothe reason I stick with X - I can throw things together quickly for X. The amount of ceremony, or big additional dependencies, needed for Wayland is ridiculous.
The only file managers that run on Wayland are the weird "flat" kind with "is" that prevent you from doing anything that didn't match their poorly conceived use cases
True.
But does not address the fact that Wayland is a bad solution to X11's problems, and that its architecturally broken from inception.