bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Van Ly via Bug reports for GNU Emacs, the Swiss army knife of text editors]

When there are three or more windows in a frame C-x o runs the command
other-window.  Wanting to goto the previous window the keybinding for
that is not obvious.  C-x p is a keymap prefix for project related work.

May I suggest to make possible the binding C-x O for (other-window -1)
as follows

(global-set-key 
 (kbd "C-x O") 
 #'(lambda ()
     "Previous window."
     (interactive)
     (other-window -1)))

Use case to demonstrate idea suggested

 1. run, emacs -Q -nw --color=no
 2. create three windows in one frame
 3. when wanting to goto the prev window, C-x o runs the command other-window that goes the long way around
 4. C-x O that runs the command (other-window -1) improves the life quality experience

The *Messages* buffer says

 C-x o (translated from C-x O) runs the command other-window

Similarly if C-j behaves like C-m in the dired buffer to run the command dired-find-file that feels better.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Eli Zaretskii]

> Date: Wed, 26 Jul 2023 06:46:30 GMT
> From:  Van Ly via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
> 
> 
> When there are three or more windows in a frame C-x o runs the command
> other-window.  Wanting to goto the previous window the keybinding for
> that is not obvious.  C-x p is a keymap prefix for project related work.

What's wrong with "C-- C-x o" (or "C-u - C-x o" if you are on a
text-mode terminal)?

> May I suggest to make possible the binding C-x O for (other-window -1)
> as follows

We don't have too many free key sequences, so "wasting" one on
something that can be easily achieved via an existing binding is not a
good idea, IMO.

Of course, you can always make this binding in your own configuration.
What I wrote above is about the default bindings -- I see no
justification to change that.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Van Ly via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> Date: Wed, 26 Jul 2023 14:15:06 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: 64868@debbugs.gnu.org
> 
> What's wrong with "C-- C-x o" (or "C-u - C-x o" if you are on a
> text-mode terminal)?
> 

C-- shrinks the font size depending on the console.

To run the universal-argument command I map the C-u to somewhere else
as the fingers more frequently use C-u for rubbing out from the
current position to left margin position zero. I use Super-u or
Escape-u for universal-argument and avoid that if I can because Super
and Escape are read differently depending on console and keyboard
context.

> > May I suggest to make possible the binding C-x O for (other-window -1)
> > as follows
> 
> We don't have too many free key sequences, so "wasting" one on
> something that can be easily achieved via an existing binding is not a
> good idea, IMO.

I agree.  I guess we could use C-x O for now until something more
preferable comes along to occupy the big O. org-mode?  That fits
nicely next to project.

> Of course, you can always make this binding in your own configuration.
> What I wrote above is about the default bindings -- I see no
> justification to change that.

Perhaps there could be a minor mode to flip the switch on C-u and C-h
behavior.  Rolling your own C-u and C-h behavior breaks functionality
for example to find completions to C-x p using

 <f1> c C-x p <f1>

doesn't do what the default C-h does I guess

 <f1> c C-x p C-h ; or TAB, I guess wrong there too

The fallback is to search the output from <f1> b or the
describe-bindings command for C-x p.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Eli Zaretskii]

> Date: Wed, 26 Jul 2023 13:19:48 GMT
> From: Van Ly <van.ly@sdf.org>
> Cc: 64868@debbugs.gnu.org
> 
> > What's wrong with "C-- C-x o" (or "C-u - C-x o" if you are on a
> > text-mode terminal)?
> 
> C-- shrinks the font size depending on the console.

In "emacs -Q"?  Doesn't happen for me.

> To run the universal-argument command I map the C-u to somewhere else
> as the fingers more frequently use C-u for rubbing out from the
> current position to left margin position zero. I use Super-u or
> Escape-u for universal-argument and avoid that if I can because Super
> and Escape are read differently depending on console and keyboard
> context.

So you will need to find a binding for (other-window -1) that is
specific to your customizations.  There's no need to change the
defaults for that, because key bindings are sort of user
customizations that users are encouraged to do to suit their needs and
preferences.

> > We don't have too many free key sequences, so "wasting" one on
> > something that can be easily achieved via an existing binding is not a
> > good idea, IMO.
> 
> I agree.  I guess we could use C-x O for now until something more
> preferable comes along to occupy the big O. org-mode?  That fits
> nicely next to project.

No, I mean we should not give this command a key binding at all, not
by default.  There's no need, since you can easily do it for you
without changing any default bindings.

> Perhaps there could be a minor mode to flip the switch on C-u and C-h
> behavior.  Rolling your own C-u and C-h behavior breaks functionality
> for example to find completions to C-x p using
> 
>  <f1> c C-x p <f1>
> 
> doesn't do what the default C-h does I guess
> 
>  <f1> c C-x p C-h ; or TAB, I guess wrong there too

Neither do "C-h c C-x p C-h", so the above is the expected behavior,
unrelated to the fact that you rebind C-h.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Van Ly via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> Date: Wed, 26 Jul 2023 16:47:52 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: 64868@debbugs.gnu.org
> 
> > Date: Wed, 26 Jul 2023 13:19:48 GMT
> > From: Van Ly <van.ly@sdf.org>
> > Cc: 64868@debbugs.gnu.org
> > 
> > > What's wrong with "C-- C-x o" (or "C-u - C-x o" if you are on a
> > > text-mode terminal)?
> > 
> > C-- shrinks the font size depending on the console.
> 
> In "emacs -Q"?  Doesn't happen for me.
> 

``emacs -Q'' inside kde/konsole has this and more discomforts for me.

> > To run the universal-argument command I map the C-u to somewhere else
> > as the fingers more frequently use C-u for rubbing out from the
> > current position to left margin position zero. I use Super-u or
> > Escape-u for universal-argument and avoid that if I can because Super
> > and Escape are read differently depending on console and keyboard
> > context.
> 
> So you will need to find a binding for (other-window -1) that is
> specific to your customizations.  There's no need to change the
> defaults for that, because key bindings are sort of user
> customizations that users are encouraged to do to suit their needs and
> preferences.
>

I found a better mapping than C-x o,

 C-x <up> for other-window in the other direction
 C-x <down> for other-window

works well with C-x <left>/<right>

A lot of mentions on the Internet say Emacs's drawback is the need to
customize it because the defaults are quirky from a different time
before keyboard commoditification.

> > > We don't have too many free key sequences, so "wasting" one on
> > > something that can be easily achieved via an existing binding is not a
> > > good idea, IMO.
> > 
> > I agree.  I guess we could use C-x O for now until something more
> > preferable comes along to occupy the big O. org-mode?  That fits
> > nicely next to project.
> 
> No, I mean we should not give this command a key binding at all, not
> by default.  There's no need, since you can easily do it for you
> without changing any default bindings.
>

Repeating several C-u -1 C-x o is suboptimal for me.

> 
> Neither do "C-h c C-x p C-h", so the above is the expected behavior,
> unrelated to the fact that you rebind C-h.
> 

The pain for the gain from remapping C-h is worth it for me, just need
f1 to function equally in place of C-h for tablike completion.

Is there a way to overlay ``b'' in the read-only modes such as
view-mode to function in the opposite direction to SPC.  The
``self-insert'' default for keys in read-only modes feels like a bug
you step on, notice but don't squish by habit.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Drew Adams]

> I found a better mapping than C-x o,
>  C-x <up> for other-window in the other direction
>  C-x <down> for other-window
> works well with C-x <left>/<right>

Again, no, please.  Bind these for yourself,
if you like.  There's no reason to add them
as default key bindings.

> A lot of mentions on the Internet say Emacs's drawback is the need to
> customize it because the defaults are quirky from a different time
> before keyboard commoditification.

Yes, well, binding C-x up/down by default,
as you suggest, just adds two more quirky
defaults.

Quirky and convenient are in the eyes of
the beholder.  Emacs lets you _easily_
get whatever key bindings you like.

Circulez. Il n'y a rien a voir.

> > we should not give this command a key binding
> > at all, not by default.  There's no need, since
> > you can easily do it for you
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > without changing any default bindings.
> 
> Repeating several C-u -1 C-x o is suboptimal for me.

"For me."  That's the point.  What's optimal
for user X isn't what user Y wants.

Fortunately, both user X and user Y can get
what they want, _trivially_.  Neither X nor
Y need to have their preferences be imposed
on others as default bindings.

"Tell Emacs" what You want, for Yourself,
and your wish is Emacs's command.  Immediate,
simple; done.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Van Ly via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> From: Drew Adams <drew.adams@oracle.com>
> CC: "64868@debbugs.gnu.org" <64868@debbugs.gnu.org>
> Thread-Topic: [External] : bug#64868: 29.1; C-x O binding suggestion for
>  (other-window -1)
> Thread-Index: AQHZxUJwcO0Jo4uOwUSm01YOR7OuKq/XDVQg
> Date: Wed, 2 Aug 2023 14:26:49 +0000
> Accept-Language: en-US
> Content-Language: en-US
> Content-Type: text/plain; charset="us-ascii"
> 
> > I found a better mapping than C-x o,
> >  C-x <up> for other-window in the other direction
> >  C-x <down> for other-window
> > works well with C-x <left>/<right>
> 
> Again, no, please.  Bind these for yourself,
> if you like.  There's no reason to add them
> as default key bindings.
> 

Well I apologise if I
suggested to change default
bindings.

I think an improvement is to
have the capability to switch
to the other window in the
other direction.  I see
keybindings like visual
themes the end user can mix
and match without too much
effort.


> > A lot of mentions on the Internet say Emacs's drawback is the need to
> > customize it because the defaults are quirky from a different time
> > before keyboard commoditification.
> 
> Yes, well, binding C-x up/down by default,
> as you suggest, just adds two more quirky
> defaults.
> 

The default is unused.

> Quirky and convenient are in the eyes of
> the beholder.  Emacs lets you _easily_
> get whatever key bindings you like.
> 
> Circulez. Il n'y a rien a voir.
> 
> > > we should not give this command a key binding
> > > at all, not by default.  There's no need, since
> > > you can easily do it for you
>     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > without changing any default bindings.
> > 
> > Repeating several C-u -1 C-x o is suboptimal for me.
> 
> "For me."  That's the point.  What's optimal
> for user X isn't what user Y wants.
> 
> Fortunately, both user X and user Y can get
> what they want, _trivially_.  Neither X nor
> Y need to have their preferences be imposed
> on others as default bindings.
> 
> "Tell Emacs" what You want, for Yourself,
> and your wish is Emacs's command.  Immediate,
> simple; done.
> 

So make it possible to toggle
in keybinding ``overlays''.

I have lived without wanting
to do the

 C-u -1 C-x o

for as long as I can remember
and put off figuring how to
make access easier.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Drew Adams]

> > Yes, well, binding C-x up/down by default,
> > as you suggest, just adds two more quirky
> > defaults.
> 
> The default is unused.

Precisely.  And the fact that it's not used
by default is no reason to use it by default.

> So make it possible to toggle in keybinding
> ``overlays''.

No idea what that means.  Whatever it is, if
you want to know how to do it, for yourself,
post a question to help-gnu-emacs@gnu.org.

> I have lived without wanting to do the
> C-u -1 C-x o
> 
> for as long as I can remember
> and put off figuring how to
> make access easier.

Just ask help-gnu-emacs@gnu.org, or some
other Emacs help venu.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Van Ly via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> From: Drew Adams <drew.adams@oracle.com>
> CC: "eliz@gnu.org" <eliz@gnu.org>,
>         "64868@debbugs.gnu.org"
> 	<64868@debbugs.gnu.org>
> Thread-Topic: [External] : bug#64868: 29.1; C-x O binding suggestion for
>  (other-window -1)
> Thread-Index: AQHZxV3sB/4gkFLMXUW4tbB7YRMCoK/XNaWg
> Date: Wed, 2 Aug 2023 16:45:36 +0000
> Accept-Language: en-US
> Content-Language: en-US
> Content-Type: text/plain; charset="us-ascii"
> 
> > > Yes, well, binding C-x up/down by default,
> > > as you suggest, just adds two more quirky
> > > defaults.
> > 
> > The default is unused.
> 
> Precisely.  And the fact that it's not used
> by default is no reason to use it by default.
> 

A wisecrack has said in the limit incompetence
and sabotage is indistinguishable.

> > So make it possible to toggle in keybinding
> > ``overlays''.
> 
> No idea what that means.

Look for toggle prefix commands for examples.

> you want to know how to do it, for yourself,
> post a question to help-gnu-emacs@gnu.org.
> 
> > I have lived without wanting to do the
> > C-u -1 C-x o
> > 
> > for as long as I can remember
> > and put off figuring how to
> > make access easier.
> 
> Just ask help-gnu-emacs@gnu.org, or some
> other Emacs help venu.
> 

Better still put it in a QFA and gradually
float popular demand to toggle-in options.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Eli Zaretskii]

> Date: Wed, 2 Aug 2023 13:07:42 GMT
> From: Van Ly <van.ly@sdf.org>
> Cc: 64868@debbugs.gnu.org
> 
> Is there a way to overlay ``b'' in the read-only modes such as
> view-mode to function in the opposite direction to SPC.

I don't understand what you mean by "overlay".  The function in the
opposite direction of SPC is View-scroll-page-backward, so if you want
to bind it to some non-default key, you can (although it already has a
convenient binding to S-SPC).

> The ``self-insert'' default for keys in read-only modes feels like a
> bug you step on, notice but don't squish by habit.

I don't understand what you are trying to say here.  What "self-insert
default" are you alluding to?

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Van Ly via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> Date: Wed, 02 Aug 2023 17:28:51 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: 64868@debbugs.gnu.org
> 
> > Date: Wed, 2 Aug 2023 13:07:42 GMT
> > From: Van Ly <van.ly@sdf.org>
> > Cc: 64868@debbugs.gnu.org
> > 
> > Is there a way to overlay ``b'' in the read-only modes such as
> > view-mode to function in the opposite direction to SPC.
> 
> I don't understand what you mean by "overlay".  The function in the
> opposite direction of SPC is View-scroll-page-backward, so if you want
> to bind it to some non-default key, you can (although it already has a
> convenient binding to S-SPC).
> 

As examples, in the occur-mode-hook and view-mode-hook I map "b" to
scroll-down-command and View-scroll-line-backward.  Can that be done
once by grouping occur-mode and view-mode as members in a read-only
mode for "b" to function that way to page in the opposite direction
SPC would scroll?

By overlay I mean how the meaning of keybindings are stacked.  At
bottom are the defaults.  A user interface lets an overlapping set to
stack above them, kind of how the visual themes work I imagine.

> > The ``self-insert'' default for keys in read-only modes feels like a
> > bug you step on, notice but don't squish by habit.
> 
> I don't understand what you are trying to say here.  What "self-insert
> default" are you alluding to?

Well, in view-mode there are many "self-insert-command" keys mapped
such as "b" or "i" or "j".  I expect the key bindings to
self-insert-command in read-only modes to be unused.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Eli Zaretskii]

> Date: Wed, 2 Aug 2023 16:45:35 GMT
> From: Van Ly <van.ly@sdf.org>
> Cc: 64868@debbugs.gnu.org
> 
> > > Is there a way to overlay ``b'' in the read-only modes such as
> > > view-mode to function in the opposite direction to SPC.
> > 
> > I don't understand what you mean by "overlay".  The function in the
> > opposite direction of SPC is View-scroll-page-backward, so if you want
> > to bind it to some non-default key, you can (although it already has a
> > convenient binding to S-SPC).
> 
> As examples, in the occur-mode-hook and view-mode-hook I map "b" to
> scroll-down-command and View-scroll-line-backward.  Can that be done
> once by grouping occur-mode and view-mode as members in a read-only
> mode for "b" to function that way to page in the opposite direction
> SPC would scroll?

Each mode has its own keymap.  So no, that's not possible, by using
just the key binding commands.

> Well, in view-mode there are many "self-insert-command" keys mapped
> such as "b" or "i" or "j".  I expect the key bindings to
> self-insert-command in read-only modes to be unused.

That's not how such Emacs modes work.  Having commands bound to
self-inserting characters make it easy to invoke those commands.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Drew Adams]

> May I suggest to make possible the binding C-x O for (other-window -1)
> as follows
> 
> (global-set-key
>  (kbd "C-x O")
>  #'(lambda ()
>      "Previous window."
>      (interactive)
>      (other-window -1)))

No, please.  Users can do that themselves, if
that's what they want.  Let's not sacrifice
another default key for such a thing.

bug#64868: 29.1; C-x O binding suggestion for (other-window -1)

[Jim Porter]

On 7/25/2023 11:46 PM, Van Ly via Bug reports for GNU Emacs, the Swiss 
army knife of text editors wrote:
> When there are three or more windows in a frame C-x o runs the command
> other-window.  Wanting to goto the previous window the keybinding for
> that is not obvious.  C-x p is a keymap prefix for project related work.
> 
> May I suggest to make possible the binding C-x O for (other-window -1)
> as follows

I believe Emacs discourages keybindings with capital letters, since it 
can cause surprising results if the user has Caps Lock on. In this case, 
Caps Lock being on would cause "C-x o" to behave "backwards".

I don't know if this is a firm rule, but we should at least take it into 
account. (Of course, there's nothing stopping you from adding this 
keybinding for your own configuration.)