I understand that RMS want a more generic approach by changing the external
representation of saved window-configurations. IMO that is overkill (and
much harder to test at this stage), so can't we just use the new function
_as_is_ (which Lennarth says is fine for his purpose), and
I sent a proposal for a new window-split-tree function to solve the
"fix bugs in balance-windows" part of this item in FOR-RELEASE:
** Make a new interface for specifying window configurations
so that we can fix bugs in balance-windows.
I understand that RMS want a more generic a
