I reworked the code to move the workspace creation to the manager to be in line with the protocol. As a result the initial state hack is gone and state/name are properly picked up from labwc. I haven't tested with a multi monitor setup, but the current version works nicely for me i.c.w. labwc. That said, I have a hard time figuring out how features like persistent-workspaces or all-ouputs are supposed to work (both functionality and code-wise). Both are very likely broken with this PR. I'm also not sure if those features still makes sense with the protocol update. May be those features should be dropped? Or may be it makes more sense to start fresh, eventually as a new module?
Leaving this for now for some feedback about the general direction.

May be those features should be dropped? Or may be it makes more sense to start fresh, eventually as a new module?

The last commit re-implements the module from scratch. I'm quite happy with the result. All config options should still work with the exception of "persistent-workspaces". I did not took this over because of above mentioned reasons.

From my perspective this is ready for testing and of course still ready for general feedback on the direction and how this should fit in.

TODO:

• Do not bump the required wayland-protocol version, but build the workspaces module conditionally based on the detected version
• Tweak documentation
• Squash commits
• Testing (limited, from developers perspective)

I'm happy with how this PR turned out and all things that needs to be done in my opinion are done. Ready for review!

The main question still remains though: should this be a new module (since the protocol is no longer wlr specific) or, what this PR is, an in-place port.

Any info on whether this will be merged?

I compiled waybar with the workspace patch and it works great with labwc, at least for my simple usecase of a desktop with a single monitor and a laptop, both with virtual desktops.