Re: Functions transpose/rotate/flip windows

[martin rudalics]

> I see, btw I think having ascii in docstrings should be fine, some other
> code in core does it, eg see docstring of
> c-ts-mode--macro-heuristic-rules (it's not ascii drawing but it's still
> diagram)

Let's concentrate on getting the drawings in the Elisp manual right first.

>>> Yes, because their is an odd number of elements.  This command would
>>> basiclly reverse the (window-list) (or the cyclic odering of windows),
>>> and in reversing, the middle pivot does not get changed if there is one.
>>>
>>>    ___________      ___________
>>> |           |    |           |
>>> |     A     |    |     D     |
>>> |___________| -> |___________|
>>> |   |   |   |    |   |   |   |
>>> | B | C | D |    | C | B | A |
>>> |___|___|___|    |___|___|___|
>>>
>>> Here the since there is even number of windows, no window is where it
>>> was before.
>>
>> And what would the new command produce instead?
>>
>
> Wdym? I put the inital layout on left and layout after command is called
> on right, I'll actually just implement it sometime to show

Above you say that "the middle pivot does not get changed if there is
one".  When I do 'rotate-windows' in the that example I get

  ___________      ___________
 |           |    |           |
 |     A     |    |     D     |
 |___________| -> |___________|
 |   |   |   |    |   |   |   |
 | B | C | D |    | A | B | C |
 |___|___|___|    |___|___|___|


So you mean the new function should behave as 'rotate-windows' with A
and C changing locations.  Right?

martin