Re: disabling undo boundaries

[Phillip Lord]

Stefan Monnier <address@hidden> writes:

>>> Then maybe you could try a hack like:
>>> 
>>> (with-current-buffer "*scratch*"
>>>   <doyourthing>)
>>> (let ((buffer-undo-list))
>>>   (undo-boundary))           ; Set last_undo_buffer.
>> It doesn't unfortunately.
>
> Oh, indeed when buffer-undo-list, undo-boundary is a no-op, so you need
>
>    (with-current-buffer "*scratch*"
>      <doyourthing>)
>    (let ((buffer-undo-list '("")))
>      (undo-boundary))           ; Set last_undo_buffer.

My fault for not being clear. AFAICT, undo-boundary does not set
last_undo_buffer at all. I can't directly test this because I can't get
to last_undo_buffer from lisp, but here is my test code.

#+BEGIN_SRC emacs-lisp
  (defvar-local fix-test-on nil)
  (defvar fix-test-fudge t)
  (defun fix-test-after-change-function (&rest _)
      (when fix-test-on
        (with-current-buffer
            (get-buffer "*scratch*")
          (insert "a"))
        (when fix-test-fudge
          (let ((buffer-undo-list '("")))
            (undo-boundary)))))

  (add-hook 'after-change-functions 'fix-test-after-change-function)
#+END_SRC

Setting `fix-test-on' to t still affects undo in the current-buffer.


>> It is, and one that has been in Emacs for a long time.  How wedded to
>> keeping it are you? Would it be possible to optionalize?
>
> Not sure,

Well I have a fork running without it (with some brutal commenting out),
so I'll try running that as my daily emacs, and see if it crashes and
how it behaves.

I'm still trying to understand the reason behind the logic in the first
place; unfortunately, the code seems to predate the earliest VC records.

Thanks for the feedback, it's been as helpful as always.

Phil