Active region corrupts undo (C-/) about lispy HOT 5 CLOSED

Can't reproduce. I get a Undo in region! message instead and
everything is fine. Could be an issue with Emacs version or a clash
with a package that I don't have.

Try to reproduce it with a minimal config:

emacs -Q -l config.el

Here's the contents of config.el (just customize your package dir):

(setq package-user-dir "~/Dropbox/source/site-lisp/elpa")
(package-initialize)
(mapc (lambda (h) (add-hook h (lambda () (lispy-mode 1))))
      '(emacs-lisp-mode-hook
        lisp-interaction-mode-hook
        clojure-mode-hook
        scheme-mode-hook
        lisp-mode-hook))
(lispy-mode 1)

from lispy.

Emacs is 24.4 compiled from source. This time with minimal config, as suggested.

(defun list-of (n elt)
  (if (zerop n)
      nil
      (cons elt (list-of (- n 1) elt))))

Run commands => q g m w s C-/

(defun list-of (n elt)
  (if (zerop n)
      nil
      (cons elt (list-of list-of elt))))

Got "Undo in region!" and resulting buffer looks like above (note: list-of list-of). Is it expected?

from lispy.

Originally I had C-/ bound to undo-tree-undo from undo-tree package required by Evil.
The binding itself presumably activated by (global-undo-tree-mode). Hope this helps.

from lispy.

OK, I can reproduce (list-of list-of elt). This is indeed a bug
which I'll try to fix soon.

As for undo-tree-undo, could you move it to a new issue?
It could be a separate bug. I haven't tried undo-tree yet, so it might take
a while to fix that one.

from lispy.

I've looked more at (list-of list-of elt). It's not a bug, since the contents
of region are restored from (- n 1) to its previous content list-of by the default
Emacs undo command.

I you want a different behavior, use lispy-undo (bound to u).
=> q g m w s u <= results in:

(defun list-of (n elt)
  (if (zerop n)
      nil
    (cons elt ((- n 1) list-of elt))))

which is reasonable, although the region is deactivated. This is because undo doesn't
track point and mark.

from lispy.