Polymode with evil-mode conflict about polymode

[vspinu]

Ok. So I am keeping same state regardless of the state. We can change that anytime. Just let me know if you have any problems.

[vsbuffalo]

Just a note, this ins't a markdown issue only:

#+begin_src R
x <- rnorm(100)
plot(density(x))
#+end_src

Editing the first line with $ a delete delete doesn't delete this text as expected.

[lionel-]

I think it's because if you are in insert state in one chunk of the file, and in normal state in another, switching from one chunk to the other will cause a change of evil state. Thus, instead of issuing your normal navigation commands, it will self insert the key you are typing.

So ideally, the evil state should be shared among all buffers displayed in a polymode document.

[vsbuffalo]

@lionelgit is there a way to tweak polymode to tell it to share evil state among buffers? This issue is quite irritating :-)

[lionel-]

there is certainly a way but someone has to write it :)

Le 24 oct. 2014 à 19:49, Vince Buffalo [email protected] a écrit :

@lionelgit https://github.com/lionelgit is there a way to tweak polymode to tell it to share evil state among buffers? This issue is quite irritating :-)

—
Reply to this email directly or view it on GitHub #41 (comment).

[vsbuffalo]

I just want to update this — there is also a very bad bug whereby evil-mode with undo-tree will absolutely lead to a destruction of your undo history when used with poly-mode. Undo will quickly insert gibberish in buffers in polymode and the lack of a working undo only makes this work. Beware!

[lionel-]

I also had this problem but didn't know it was caused by undo-tree. I lost data a few times. Thanks for the information!

[vsbuffalo]

Is there any update on this?

[vspinu]

Not really. Sorry. There are too many non-trivial issues to deal with and I haven't had time to get into it as yet. But I will surely do, sooner or later.

[holtzermann17]

Although it's for a different package, it looks like this is the idea: https://github.com/tarao/multi-mode-util/blob/master/multi-mode%2Bevil.el

[lionel-]

it would be very nice for someone to write such a package, tough I don't have time to write one in the foreseeable future.

This package could also provide the following text objects for chunks:

  (defun polymode-mark-inner-chunk ()
    (let ((span (pm/get-innermost-span)))
      (set-mark (1+ (nth 1 span)))
      (goto-char (1- (nth 2 span)))))

  (defun polymode-mark-chunk ()
    (let ((span (pm/get-innermost-span)))
      (goto-char (nth 1 span))
      (set-mark (line-beginning-position))
      (goto-char (nth 2 span))
      (goto-char (line-end-position))))

  (evil-define-text-object evil-inner-polymode-chunk (count &optional beg end type)
    "Select inner polymode chunk."
    :type line
    (polymode-mark-inner-chunk)
    (evil-range (region-beginning)
                (region-end)
                'line))

  (evil-define-text-object evil-a-polymode-chunk (count &optional beg end type)
    "Select a polymode chunk."
    :type line
    (polymode-mark-chunk)
    (evil-range (region-beginning)
                (region-end)
                'line))

[vspinu]

I have finally added integration for evil mode. Sorry for taking so long.

My understanding of modal editing hardly goes beyond hjkl but it seems to work as expected. I would very much like to hear form an evil user that it actually works.

Inspired by the multi-mode-evil.el I have introduced a custom variable:

(defcustom polymode-evil-states '(normal insert emacs)
  "States of evil-mode to be preserved when switching indirect buffers."
  :type '(symbol)
  :group 'polymode-compat)

Is this really needed? Don't you want to preserve all the states when moving between spans and chunks?

[holtzermann17]

On Sun, Feb 28, 2016 at 1:39 AM, Vitalie Spinu [email protected]
wrote:

Don't you want to preserve all the states when moving between spans and
chunks?

That policy makes sense to me.

(I'll try to test the changed mode soon.)

[roryk]

@vspinu, this is AWESOME.

[stefanavey]

I'm having the problem referenced above by @vsbuffalo where using undo-tree minor mode within polymode will insert gibberish into buffer on occasion and the undo history seems corrupted. I'm not using evil-mode so I don't think it's related to that but wondering if it was resolved for you.

[philipphoman]

Same here, using evil-mode. Undo history corrupted.

[vspinu]

Undo movement between indirect buffers is handled by emacs internals. If undo is corrupted that could mean something is not quite right other there. IN any case I would apreciate a reproducible example here. Would you mind opening a new issue for this?

[jody-frankowski]

This should be a reproducible example: #214