First cut at a cross-"platform" X11/Wayland windowing library

[Brian Tarricone]

On Fri, Sep 23, 2022, at 15:32, adlo wrote:

> I think the ext-workspace protocol could be modified to announce what 
> zwlr_foreign_toplevel_handles are on what workspace. An ext_workspace 
> could announce when a window enters it. I don't think that would 
> require the cooperation of the wlr_foreign_toplevel devs, because the 
> compositor would know internally which zwlr_foreign_toplevel_handles 
> are on which ext_workspace. But I think that's the wrong way of doing 
> it, as that would be duplicating wlr_foreign_toplevel's "toplevel" 
> event, and I think those events more properly belong in the 
> wlr_foreign_toplevel protocol, not the ext_workspace protocol.
>
> Is it that the wlr_foreign_toplevel protocol would require some new 
> "workspace-enter" and "workspace-leave" events?

I'm still new to the whole Wayland protocol thing, so I don't know that I have a great grasp on what is and isn't a good best practice when two protocols need to interact like that.  Interesting question of which of the two protocols should be responsible for announcing attributes that relate to the other.

But yeah, either having the foreign-toplevel protocol send workspace-enter and workspace-leave events, or having the workspace protocol send toplevel-enter and toplevel-leave events, would work.  Conceptually, though, I'm not sure if it makes more sense for the foreign-toplevel protocol to "know about" the workspace concept, or for the workspace protocol to "know about" toplevels.  I do kinda feel like standalone foreign toplevels as a concept is perfectly good on its own.  And the entire point of workspaces is to allow you to organize your toplevel windows, so maybe toplevel-enter and toplevel-leave events on the workspace protocol makes more sense.  But I don't think I'd object to the opposite proposal, *shrug*.

     -brian