Re: Window state/configuration (and markers)

Re: Window state/configuration (and markers)

[Emanuel Berg]

Carlos Pita <carlosjosepita@gmail.com> writes:

> I would like to save/restore window
> state/configuration ...

What problem or undesirable situation do you
experience?

I seldom find it sensible to store and load window
configurations. But if you habitually setup very
ambitious configs with debugging, documentation (man
pages, browser, etc.), source, some e-mail or post
discussion the source, the file directory, and
everything else you can think of... yeah.

This is the material I have on this. It works but I
don't know about the mark or everything else you
mention.

(setq default-win-config-reg ?0)

(defun save-win-config (&optional reg)
   (interactive "P")
   (window-configuration-to-register
    (if reg reg default-win-config-reg) ))

(defun fetch-win-config (&optional reg)
  (interactive "P")
  (jump-to-register
   (if reg reg default-win-config-reg)) )

-- 
underground experts united

Window state/configuration (and markers)

[Carlos Pita]

Hi all,

I would like to save/restore window state/configuration without
including markers (point, start). I'm aware I'm neither the only nor
the first one looking for this and I know of a number of hacks and
advices to get it working. That's not my issue, then.

Instead I would like to focus here in the relationship between
window-configurations (opaque type) and window-states (relatively new
elisp-level alternative) and the reason why the latter seems to be
poorly adopted by the myriad of
workspaces/workgroups/perspectives/winmgrs extensions and why it's not
supported by vanilla emacs to the same extent than the former.

So:

1) Do you know of some extension already using the window-state machinery?

2) Why no excursion, registry, etc. functionality is provided
out-of-the-box as a counterpart to the window-configuration ones?

3) Are the window-state-get/set functions the current way to go or
just another tool in the toolset? Are window-configurations kept for
historical reasons or have they some technical advantages over
window-states?

Finally, this one is specifically related to my no-marker requirement:

4) Years ago there was some discussion in the dev list about adding an
option to avoid saving markers in window-configurations. Someone
pointed out that the window-state-* functions were just to be merged
and the discussion then diverted to this more exciting topic. Nowadays
I'm not able to find any quick option to get the no-marker
configuration. True, window-state-get has the writable argument, but
the out-of-the-box registry functions and mappings are still missing.
Do you think it would be of some worth to have a more ready to use
solution here? I mean, the save/restore layout use case is best served
by a no-marker configuration/state.

Best regards
--
Carlos