Hello Tony,
thanks for your detailed explanation.
> So this is very expressive, but a bit cryptic and not very well
> documented outside of the testcases. It also plays to the primary
> strength of the workspace reply object which is that no recursive
> tree-parsing is required so it can be used with simple tools.
The i3bar workspace parsing code still has to be adjusted because of its
very strict nature. I was quite surprised that there was no i3bar when I
first started i3 after my changes!
> Would this implementation fit your use case?
Yes, that would fit my use case perfectly well.
I was essentially doing the same thing, just using nodes_head and
floating_head instead. But after adjusting the code to use focus_head,
several test cases fail.
Problematic seems to be that I have to alter focus_head before the
window close event is send, in order to get the correct window count. It
worked just fine with the other mentioned data structures. I will try
and see whether I can get it working again. If I can't, I will still
post my working patch and we can go from there. Thanks.
Cheers,
Marco