bug#29050: 26.0; Change in when `window-configuration-change-hook' is run

bug#29050: 26.0; Change in when `window-configuration-change-hook' is run

[Drew Adams]

This change does not seem right (from NEWS):

 *** Resizing a frame no longer runs 'window-configuration-change-hook'.
 'window-size-change-functions' should be used instead.

Previously you could have a hook on 'window-configuration-change-hook'
that would take effect for frame resizings.  Now you cannot.

Perhaps someone thought that just telling users to use
'window-size-change-functions' instead would suffice.  No.  That hook
(which already existed, and which was fine as it was) is for ABNORMAL
hooks.  This incompatible change means that you cannot use the same,
NORMAL hook for both'window-configuration-change-hook' and
'window-size-change-functions'.

So if you want the behavior you had before, i.e., you want a function to
be invoked for both kinds of changes, you are out of luck.  You need to
have two different functions, or you need to at least change the
function to accept a frame argument, even if it is not used.  Why?

Example:

(define-minor-mode pretty-control-l-mode
    "Toggle pretty display of Control-l (`^L') characters.
With ARG, turn pretty display of `^L' on if and only if ARG is positive."
  :init-value nil :global t :group 'Pretty-Control-L
  (if pretty-control-l-mode
      (add-hook 'window-configuration-change-hook 'refresh-pretty-control-l)
    (remove-hook 'window-configuration-change-hook 'refresh-pretty-control-l))
  (walk-windows
   (lambda (window)
     (let ((display-table  (or (window-display-table window)
                               (make-display-table))))
       (aset display-table ?\014 (and pretty-control-l-mode
                                      (pp^L-^L-display-table-entry window)))
       (set-window-display-table window display-table)))
   'no-minibuf
   'visible))

The hook function no longer kicks in for "frame resizing", which also
means that it no longer kicks in when a frame is created.  So now the
code needs to add the hook function to both hooks (a normal hook and an
abnormal hook).  And the hook function, `refresh-pretty-control-l', now
needs to be changed to accept a phantom FRAME arg:

(defun refresh-pretty-control-l (&optional _)
  "Reinitialize `pretty-control-l-mode', if on, to update the display."
  (interactive)
  (when pretty-control-l-mode (pretty-control-l-mode t)))

Why this incompatible change?

In GNU Emacs 26.0.90 (build 3, x86_64-w64-mingw32)
 of 2017-10-13
Repository revision: 906224eba147bdfc0514090064e8e8f53160f1d4
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --without-dbus --host=x86_64-w64-mingw32
 --without-compress-install 'CFLAGS=-O2 -static -g3''

bug#29050: 26.0; Change in when `window-configuration-change-hook' is run

[martin rudalics]

 >   *** Resizing a frame no longer runs 'window-configuration-change-hook'.
 >   'window-size-change-functions' should be used instead.
 >
 > Previously you could have a hook on 'window-configuration-change-hook'
 > that would take effect for frame resizings.  Now you cannot.
 >
 > Perhaps someone thought that just telling users to use
 > 'window-size-change-functions' instead would suffice.  No.  That hook
 > (which already existed, and which was fine as it was) is for ABNORMAL
 > hooks.  This incompatible change means that you cannot use the same,
 > NORMAL hook for both'window-configuration-change-hook' and
 > 'window-size-change-functions'.
 >
 > So if you want the behavior you had before, i.e., you want a function to
 > be invoked for both kinds of changes, you are out of luck.  You need to
 > have two different functions, or you need to at least change the
 > function to accept a frame argument, even if it is not used.  Why?

In earlier Emacsen, changing the frame size was allowed to delete all
windows but the selected one.  Nowadays, changing the frame size does
not delete windows any more and so does no more change the window
configuration proper but only the sizes of individual windows.

'window-configuration-change-hook' is a quite expensive hook which is
run for way too many functions in the windows area.  It's expensive
because it's run for every single invocation of these functions and
often even if nothing had changed at all.  I plan to run it exclusively
for the following cases - window deletion and creation and displaying a
different buffer in a window - which are all real changes.  Any window
resizing is covered by ‘window-size-change-functions’ which is now run
(almost) only when the size of a window really changed.

martin

bug#29050: 26.0; Change in when `window-configuration-change-hook' is run

[Drew Adams]

> In earlier Emacsen, changing the frame size was allowed to delete all
> windows but the selected one.  Nowadays, changing the frame size does
> not delete windows any more and so does no more change the window
> configuration proper but only the sizes of individual windows.
> 
> 'window-configuration-change-hook' is a quite expensive hook which is
> run for way too many functions in the windows area.  It's expensive
> because it's run for every single invocation of these functions and
> often even if nothing had changed at all.  I plan to run it exclusively
> for the following cases - window deletion and creation and displaying a
> different buffer in a window - which are all real changes.  Any window
> resizing is covered by ‘window-size-change-functions’ which is now run
> (almost) only when the size of a window really changed.

OK.

bug#29050: 26.0; Change in when `window-configuration-change-hook' is run

[martin rudalics]

 >> 'window-configuration-change-hook' is a quite expensive hook which is
 >> run for way too many functions in the windows area.  It's expensive
 >> because it's run for every single invocation of these functions and
 >> often even if nothing had changed at all.  I plan to run it exclusively
 >> for the following cases - window deletion and creation and displaying a
 >> different buffer in a window - which are all real changes.  Any window
 >> resizing is covered by ‘window-size-change-functions’ which is now run
 >> (almost) only when the size of a window really changed.
 >
 > OK.

Closing this bug then.

Thanks, martin