bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Stefan Kangas]

Severity: wishlist

Please consider adding the following keybindings to outline-mode:

(define-key outline-minor-mode-map (kbd "M-<left>") 'outline-demote)
(define-key outline-minor-mode-map (kbd "M-<right>") 'outline-promote)
(define-key outline-minor-mode-map (kbd "M-<up>") 'outline-move-subtree-up)
(define-key outline-minor-mode-map (kbd "M-<down>") 'outline-move-subtree-down)

They are already there in Org-mode and many users swear by them.

Background:
https://lists.gnu.org/archive/html/emacs-devel/2020-04/msg01887.html

These keybindings were first suggested by Howard Melman in the above
thread.

Best regards,
Stefan Kangas

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Stefan Kangas]

> (define-key outline-minor-mode-map (kbd "M-<left>") 'outline-demote)
> (define-key outline-minor-mode-map (kbd "M-<right>") 'outline-promote)

Sorry, that should be the other way around:

(define-key outline-minor-mode-map (kbd "M-<left>") 'outline-promote)
(define-key outline-minor-mode-map (kbd "M-<right>") 'outline-demote)

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Lars Ingebrigtsen]

Stefan Kangas <stefan@marxist.se> writes:

>> (define-key outline-minor-mode-map (kbd "M-<left>") 'outline-demote)
>> (define-key outline-minor-mode-map (kbd "M-<right>") 'outline-promote)
>
> Sorry, that should be the other way around:
>
> (define-key outline-minor-mode-map (kbd "M-<left>") 'outline-promote)
> (define-key outline-minor-mode-map (kbd "M-<right>") 'outline-demote)

`M-<right>' (etc) are already bound to cursor movement commands by
default, so this would shadow these bindings in outline-minor-mode?  I
think that would be inconvenient for many people.

I'm not sure M-<right> (etc) are obvious bindings for outline handling
even if that weren't the case.  Anybody got any opinions here?

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

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Pankaj Jangid]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>>> (define-key outline-minor-mode-map (kbd "M-<left>") 'outline-demote)
>>> (define-key outline-minor-mode-map (kbd "M-<right>") 'outline-promote)
>>
>> Sorry, that should be the other way around:
>>
>> (define-key outline-minor-mode-map (kbd "M-<left>") 'outline-promote)
>> (define-key outline-minor-mode-map (kbd "M-<right>") 'outline-demote)
>
> `M-<right>' (etc) are already bound to cursor movement commands by
> default, so this would shadow these bindings in outline-minor-mode?  I
> think that would be inconvenient for many people.
>
> I'm not sure M-<right> (etc) are obvious bindings for outline handling
> even if that weren't the case.  Anybody got any opinions here?

That is difficult question. Because people coming from org background
will find it difficult if ‘M-<right>’ and ‘M-<left>’ don’t work as
expected by them. But then how many people will use both, org and the
outline-mode.

Irrespective of the above issue, in my opinion, there should be a
standardized way of doing similar things in Emacs.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Stefan Kangas]

Pankaj Jangid <pankaj@codeisgreat.org> writes:

> Lars Ingebrigtsen <larsi@gnus.org> writes:
>
>>> (define-key outline-minor-mode-map (kbd "M-<left>") 'outline-promote)
>>> (define-key outline-minor-mode-map (kbd "M-<right>") 'outline-demote)
>>
>> `M-<right>' (etc) are already bound to cursor movement commands by
>> default, so this would shadow these bindings in outline-minor-mode?  I
>> think that would be inconvenient for many people.

Yes, it is a real problem.  But since this is context dependent, I find
that in practice I am rarely surprised by this.  YMMV.

>> I'm not sure M-<right> (etc) are obvious bindings for outline handling
>> even if that weren't the case.  Anybody got any opinions here?
>
> That is difficult question. Because people coming from org background
> will find it difficult if ‘M-<right>’ and ‘M-<left>’ don’t work as
> expected by them.

Right, this is the reason for suggesting this in the first place.

FWIW, I find the org-mode bindings to be okay in the sense that at least
they are easy to learn.  But I have found myself wanting something a bit
more ergonomic at times.

> Irrespective of the above issue, in my opinion, there should be a
> standardized way of doing similar things in Emacs.

Agreed.  Perhaps it is possible to find a better alternative.  Such an
alternative should ideally be suitable for at least all three of
outline-mode, org-mode and outline-minor-mode.

It might be hard to settle on something that makes everyone happy,
however.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Lars Ingebrigtsen]

Stefan Kangas <stefan@marxist.se> writes:

>> Irrespective of the above issue, in my opinion, there should be a
>> standardized way of doing similar things in Emacs.
>
> Agreed.  Perhaps it is possible to find a better alternative.  Such an
> alternative should ideally be suitable for at least all three of
> outline-mode, org-mode and outline-minor-mode.
>
> It might be hard to settle on something that makes everyone happy,
> however.

Indeed.  Major modes (like Org) have greater leeway in defining
keystrokes than minor modes, though -- a minor mode has to be more
careful not to redefine common keystrokes because there's less of a "buy
in" from users than with a major mode like Org.

So it might make sense for Org to bind M-<right> to something while it'd
be too intrusive for outline-minor-mode.

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

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Howard Melman]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Stefan Kangas <stefan@marxist.se> writes:
>
>>> Irrespective of the above issue, in my opinion, there should be a
>>> standardized way of doing similar things in Emacs.
>>
>> Agreed.  Perhaps it is possible to find a better alternative.  Such an
>> alternative should ideally be suitable for at least all three of
>> outline-mode, org-mode and outline-minor-mode.
>>
>> It might be hard to settle on something that makes everyone happy,
>> however.
>
> Indeed.  Major modes (like Org) have greater leeway in defining
> keystrokes than minor modes, though -- a minor mode has to be more
> careful not to redefine common keystrokes because there's less of a "buy
> in" from users than with a major mode like Org.
>
> So it might make sense for Org to bind M-<right> to something while it'd
> be too intrusive for outline-minor-mode.

If I'm correct that the standard cursor movement bindings on
M-<right>, M-<left> you're referring to are word movement
commands, then they also have existing bindings on C-<right>
and C-<left> and are at least very similar to M-f and M-b..
So perhaps it's ok if outline-minor-mode shadows the
M-<arrow> flavor of them, particularly if it aligns with
org-mode.

-- 

Howard

bug#41129: [External] : bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Drew Adams]

> > Indeed.  Major modes (like Org) have greater leeway in defining
> > keystrokes than minor modes, though -- a minor mode has to be more
> > careful not to redefine common keystrokes because there's less of a
> > "buy in" from users than with a major mode like Org.
> >
> > So it might make sense for Org to bind M-<right>
> > to something while it'd be too intrusive for
> > outline-minor-mode.

I agree with Lars, there.

A minor mode, even when local to a buffer (not
explicitly global), is in a sense "global" - it
works across other modes, and needs to play well
with them.
___

Not directly related, but I also think it's a
no-no for a major mode such as Org to globally
bind a key that has no business being used
(because useless) outside Org mode.  That's the
point of bug #46087 (which has received no reply).

In the past, Org mode has even countered the key
binding guidelines, by binding keys in its major
mode map that are supposed to be reserved for
minor modes.  (I think that after some protest
that's now been corrected.)

I can understand that Org developers and users
would like many keys available by default, but
there's also a need to play well with other,
non-Org, uses of Emacs.

> If I'm correct that the standard cursor movement bindings on
> M-<right>, M-<left> you're referring to are word movement
> commands, then they also have existing bindings on C-<right>
> and C-<left> and are at least very similar to M-f and M-b.

Yes.

> So perhaps it's ok if outline-minor-mode shadows the
> M-<arrow> flavor of them, particularly if it aligns with
> org-mode.

Personally, I'd prefer that Emacs not bind
`M-left|right' keys by default, including in
Outline or Org modes (minor or major).

(I bind `M-left|right|up|down' to commands
that incrementally move the selected frame.)

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Lars Ingebrigtsen]

Howard Melman <hmelman@gmail.com> writes:

> If I'm correct that the standard cursor movement bindings on
> M-<right>, M-<left> you're referring to are word movement
> commands, then they also have existing bindings on C-<right>
> and C-<left> and are at least very similar to M-f and M-b..

That these commands also have other bindings doesn't really affect how
annoying rebinding M-<right> is for the users that are used to M-<right>.

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

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Howard Melman]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Howard Melman <hmelman@gmail.com> writes:
>
>> If I'm correct that the standard cursor movement bindings on
>> M-<right>, M-<left> you're referring to are word movement
>> commands, then they also have existing bindings on C-<right>
>> and C-<left> and are at least very similar to M-f and M-b..
>
> That these commands also have other bindings doesn't really affect how
> annoying rebinding M-<right> is for the users that are used to M-<right>.

So I'm clear, the command right-word, which is only useful
when editing bidirectional (not merely right-to-left) text ,
and even then only changes the orientation of the very basic
command forward-word (still bound to M-f), is so important
it needs TWO very similar default key bindings (M-right and
C-right).  And an optional minor-mode, when enabled, can't
shadow ONE of those bindings so that its very commonly used
command outline-demote can be, by default, on the easy to
type and remember M-right instead of only the difficult to
type and remember C-c @ C-right to match the behavior of the
very popular org-mode.

I agree that default bindings should be changed with
caution, even great caution, but IMHO if this case doesn't
meet that threshold I'm hard pressed to think of another
that can.

-- 

Howard

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Eli Zaretskii]

> From: Howard Melman <hmelman@gmail.com>
> Date: Sat, 30 Jan 2021 12:39:52 -0500
> 
> the command right-word, which is only useful when editing
> bidirectional

This assertion is incorrect.  The real goal of right-word is to
provide a command that works as users expect in both left-to-right and
right-to-left contexts.  The alternative would be to request that the
user should decide whether to use M-f or M-b to go in a particular
direction, and that's a non-starter.

> is so important it needs TWO very similar default key bindings
> (M-right and C-right)

This command has two bindings not because of its importance, but
because other programs out there use those bindings.  Emacs is
following the expectations of the users here.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Howard Melman]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Howard Melman <hmelman@gmail.com>
>> Date: Sat, 30 Jan 2021 12:39:52 -0500
>> 
>> the command right-word, which is only useful when editing
>> bidirectional
>
> This assertion is incorrect.  The real goal of right-word is to
> provide a command that works as users expect in both left-to-right and
> right-to-left contexts.  The alternative would be to request that the
> user should decide whether to use M-f or M-b to go in a particular
> direction, and that's a non-starter.

I stand corrected.  I'm a left-to-right only user and I
won't make the case that that is more important, merely that
I'm unfamiliar with the specifics of other users.

>> is so important it needs TWO very similar default key bindings
>> (M-right and C-right)
>
> This command has two bindings not because of its importance, but
> because other programs out there use those bindings.  Emacs is
> following the expectations of the users here.

Ok, but in emacs they expect and need both?  How do they
adjust in org mode? Or is this a complaint that users of
right-word have in org-mode? Would users of
outline-minor-mode adjust similarly?

Does outline-mode support right-to-left oriented outlines?
Does org?  I'm guessing not because then these arrow key
bindings would want to adjust demote/promote behavior
accordingly.  Should they?

Maybe these bindings could be in another minor-mode that
could be enabled by those users that want them?
outline-use-org-bindings-minor-mode. 

-- 

Howard

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Eli Zaretskii]

> From: Howard Melman <hmelman@gmail.com>
> Date: Sat, 30 Jan 2021 13:48:38 -0500
> 
> > This command has two bindings not because of its importance, but
> > because other programs out there use those bindings.  Emacs is
> > following the expectations of the users here.
> 
> Ok, but in emacs they expect and need both?

Yes, because some applications support one and others the other key
binding.

> How do they adjust in org mode? Or is this a complaint that users of
> right-word have in org-mode? Would users of outline-minor-mode
> adjust similarly?

The bindings have nothing to do with bidirectional text.  C-<RIGHT>
and C-<LEFT> (and the corresponding M- bindings) have a meaning in
strict left-to-right text as well: they move by words in the
corresponding directions.

> Does outline-mode support right-to-left oriented outlines?
> Does org?

Yes to both questions, but again: the issue at hand has nothing to do
with bidirectional editing.

> I'm guessing not because then these arrow key bindings would want to
> adjust demote/promote behavior accordingly.  Should they?

That's a separate issue, for which I have no definite opinions to
offer.  I chimed in to set the record straight on these bindings
regardless of Org or Outline.  Whether rebinding these to
level-related commands will or will not confuse users of Org and
Outline is something for those users to say.

> Maybe these bindings could be in another minor-mode that
> could be enabled by those users that want them?
> outline-use-org-bindings-minor-mode. 

I think the idea is that many users want them.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Howard Melman]

Eli Zaretskii <eliz@gnu.org> writes:

> The bindings have nothing to do with bidirectional text.  C-<RIGHT>
> and C-<LEFT> (and the corresponding M- bindings) have a meaning in
> strict left-to-right text as well: they move by words in the
> corresponding directions.

>> Does outline-mode support right-to-left oriented outlines?
>> Does org?
>
> Yes to both questions, but again: the issue at hand has nothing to do
> with bidirectional editing.

I mean yes, they do something expected in left-to-right
text, but in strict left-to-right usage these bindings are
unnecessary because of M-f and M-b.  It's not like emacs
enables cua-mode by default because other programs use those
bindings and users expect it.

Even the emacs manual says:  "This command (‘right-word’)
behaves like ‘M-f’, except it moves _backward_ by one word
if the current paragraph is right-to-left. *Note Bidirectional Editing."

>> Maybe these bindings could be in another minor-mode that
>> could be enabled by those users that want them?
>> outline-use-org-bindings-minor-mode. 
>
> I think the idea is that many users want them.

I agree.  

-- 

Howard

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Eli Zaretskii]

> From: Howard Melman <hmelman@gmail.com>
> Date: Sat, 30 Jan 2021 14:40:37 -0500
> 
> I mean yes, they do something expected in left-to-right
> text, but in strict left-to-right usage these bindings are
> unnecessary because of M-f and M-b.

They are necessary because users (including users of strict
left-to-right text) expect C-<RIGHT> etc. to do what they do in
Emacs.  M-f and M-b are Emacs-specific.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Juri Linkov]

> Maybe these bindings could be in another minor-mode that
> could be enabled by those users that want them?
> outline-use-org-bindings-minor-mode.

bug#45147 implemented new mode outline-cycle-minor-mode
for Org-like TAB cycling in outline-minor-mode.

It's easy to customize outline-mode-cycle-map
and to bind M-left/right to outline-promote/demote:

(define-key outline-mode-cycle-map [(meta left)] 'outline-promote)
(define-key outline-mode-cycle-map [(meta right)] 'outline-demote)

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Lars Ingebrigtsen]

Juri Linkov <juri@linkov.net> writes:

> bug#45147 implemented new mode outline-cycle-minor-mode
> for Org-like TAB cycling in outline-minor-mode.
>
> It's easy to customize outline-mode-cycle-map
> and to bind M-left/right to outline-promote/demote:
>
> (define-key outline-mode-cycle-map [(meta left)] 'outline-promote)
> (define-key outline-mode-cycle-map [(meta right)] 'outline-demote)

It's now outline-minor-mode-cycle-map, but I added the four bindings
suggested by Stefan to Emacs 29 now, and I guess we'll find out whether
that gets in the way of people's work flows here.

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

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Howard Melman]

> On May 9, 2022, at 6:38 AM, Lars Ingebrigtsen <larsi@gnus.org> wrote:
> 
> Juri Linkov <juri@linkov.net> writes:
> 
>> bug#45147 implemented new mode outline-cycle-minor-mode
>> for Org-like TAB cycling in outline-minor-mode.
>> 
>> It's easy to customize outline-mode-cycle-map
>> and to bind M-left/right to outline-promote/demote:
>> 
>> (define-key outline-mode-cycle-map [(meta left)] 'outline-promote)
>> (define-key outline-mode-cycle-map [(meta right)] 'outline-demote)
> 
> It's now outline-minor-mode-cycle-map, but I added the four bindings
> suggested by Stefan to Emacs 29 now, and I guess we'll find out whether
> that gets in the way of people's work flows here.

Should these four bindings be on outline-minor-mode-cycle-map or
on outline-minor-mode-map?  They don't have to do with cycling.

Howard

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Lars Ingebrigtsen]

Howard Melman <hmelman@gmail.com> writes:

> Should these four bindings be on outline-minor-mode-cycle-map or
> on outline-minor-mode-map?  They don't have to do with cycling.

Juri suggested putting them on the former because they'd be too
disturbing in the main map -- M-<right> etc are bound to navigation
commands in editable buffers.

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

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Juri Linkov]

>> Should these four bindings be on outline-minor-mode-cycle-map or
>> on outline-minor-mode-map?  They don't have to do with cycling.
>
> Juri suggested putting them on the former because they'd be too
> disturbing in the main map -- M-<right> etc are bound to navigation
> commands in editable buffers.

Actually I suggested these bindings only for personal customization
for anyone who needs to edit heading levels because these bindings
are too intrusive to be used by default in the cycling map.
It already takes the TAB key from users, and taking more keys is too much
for the cycling map.  To be able to add editing keys to outline-minor-mode
they need to be on some separate map such as outline-minor-mode-edit-map.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Visuwesh]

[புதன் மே 11, 2022] Juri Linkov wrote:

>>> Should these four bindings be on outline-minor-mode-cycle-map or
>>> on outline-minor-mode-map?  They don't have to do with cycling.
>>
>> Juri suggested putting them on the former because they'd be too
>> disturbing in the main map -- M-<right> etc are bound to navigation
>> commands in editable buffers.
>
> Actually I suggested these bindings only for personal customization
> for anyone who needs to edit heading levels because these bindings
> are too intrusive to be used by default in the cycling map.
> It already takes the TAB key from users, and taking more keys is too much
> for the cycling map.  To be able to add editing keys to outline-minor-mode
> they need to be on some separate map such as outline-minor-mode-edit-map.

Moreover, with repeat-mode turned on, changing the outline level and
moving them is not tedious anymore.

bug#41129: outline-mode: New keybindings for demote/promote/move-subtree

[Lars Ingebrigtsen]

Juri Linkov <juri@linkov.net> writes:

> Actually I suggested these bindings only for personal customization
> for anyone who needs to edit heading levels because these bindings
> are too intrusive to be used by default in the cycling map.
> It already takes the TAB key from users, and taking more keys is too much
> for the cycling map.  To be able to add editing keys to outline-minor-mode
> they need to be on some separate map such as outline-minor-mode-edit-map.

Sorry; I misunderstood you.  I've now reverted the change.

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