bug#23903: 25.0.95; Undo after keyboard shortcuts differs from undo after M-x ... RET

bug#23903: 25.0.95; Undo after keyboard shortcuts differs from undo after M-x ... RET

[Markus Triska]

With the latest emacs-25 git, please do the following:

1) Invoke Emacs via:

    emacs -Q --eval "(defun bc () (interactive)\
                         (start-process \"bc\" (current-buffer) \"bc\"))" \
          --eval "(global-set-key [f9] 'bc)" \
          --eval "(save-excursion (insert \"hello\"))"

2) Press (F9 is bound to the function bc by the snippet above):

   F9 C-/

   After undo, point is at the beginning of the buffer, as expected.

3) Press:

   M-x bc RET C-/

   After undo, point is unexpectedly at the end of the buffer instead.

Thank you for looking into this!

All the best,
Markus


In GNU Emacs 25.0.95.4 (x86_64-apple-darwin15.5.0, X toolkit, Xaw scroll bars)
 of 2016-07-05 built on mt-imac
Repository revision: e3b039d1a0e611d6619ed3ce67d125160d644ebc
Windowing system distributor 'The X.Org Foundation', version 11.0.11502000
Configured using:
 'configure --without-ns CFLAGS=-I/opt/local/include
 LDFLAGS=-L/opt/local/lib'

Configured features:
XPM JPEG TIFF GIF PNG RSVG IMAGEMAGICK GSETTINGS NOTIFY ACL GNUTLS
LIBXML2 FREETYPE XFT ZLIB TOOLKIT_SCROLL_BARS LUCID X11

Important settings:
  value of $LANG: en_US.UTF-8
  locale-coding-system: utf-8-unix

bug#23903: 25.0.95; Undo after keyboard shortcuts differs from undo after M-x ... RET

[Lars Ingebrigtsen]

Markus Triska <triska@metalevel.at> writes:

> With the latest emacs-25 git, please do the following:
>
> 1) Invoke Emacs via:
>
>     emacs -Q --eval "(defun bc () (interactive)\
>                          (start-process \"bc\" (current-buffer) \"bc\"))" \
>           --eval "(global-set-key [f9] 'bc)" \
>           --eval "(save-excursion (insert \"hello\"))"
>
> 2) Press (F9 is bound to the function bc by the snippet above):
>
>    F9 C-/
>
>    After undo, point is at the beginning of the buffer, as expected.
>
> 3) Press:
>
>    M-x bc RET C-/
>
>    After undo, point is unexpectedly at the end of the buffer instead.

(This bug report unfortunately got no response at the time.)

Thanks for a clear recipe for reproducing this bug.

I can confirm that this behaviour is still present in Emacs 28.  Does
anybody have any insight into why this may be happening (before I start
trying to debug it)?  

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no