I’m sure there’s good reasons for this design decision but this overall situation is why Linux display will never hold a candle to Windows or MacOS or iOS. This kind of thing should just be monolithic, if not architecture then in politics. The people writing all the different components should answer to the same unit tests, the same specifications, and the same boss. A truly great solution needs unity of vision and unity of design.

And dbus for things that aren't in-scope for Wayland itself.

That's kinda the full issue in standardizing. Gnome wants dbus for everything while sway wants wayland extension most things as they avoid using dbus code. So stuff like xsettings/xrdb or screensharing replacement has basically has wayland extension and a dbus API.

It looks like thanks to Wayland being more modular, there is even more room for fragmentation. X had its fair share of extensions, but somehow most WMs and DMs managed to agree on things.

And in my opinion the whole isolation and security concept is hot garbage. It hinders so many useful things. My Linux desktop is not a smartphone where I download random, badly screened closed source apps from a play store. I'm downloading open source tools via my distro's package manager. If one of these got backdoored, Wayland preventing it from taking a screenshot of another Wayland app won't exactly save the day anyways.

Instead we're now getting clumsy, overly complicated solutions for all these simple use cases. Ultimately I don't care. If people enjoy creating these needlessly complicated monstrosities, fine. It's just that we need to wait ten times as long until we get something usable that way. I guess X needs to keep chugging along a couple more years....

How did any of that JavaScript have access to anything that would let it affect X?

> I guess X needs to keep chugging along a couple more years....

Likely at least a decade (Lindy effect).