keyboard-translate working for Control but not for Alt

keyboard-translate working for Control but not for Alt

[Elena]

Hello,

since I use a Dvorak keyboard, I'd like to swap C-x and C-t, as it is
suggested here:

http://www.emacswiki.org/emacs/DvorakKeyboard#toc4

The proposed code performs its task:

http://www.emacswiki.org/emacs/DvorakKeyboard#toc4

(keyboard-translate ?\C-x ?\C-t)
(keyboard-translate ?\C-t ?\C-x)

Out of curiosity, I tried to swap M-x and M-t as well:

(keyboard-translate ?\M-x ?\M-t)
(keyboard-translate ?\M-t ?\M-x)

but evaluating each expression fail with (numbers are not the same):

Debugger entered--Lisp error: (wrong-type-argument characterp
134217848)
  keyboard-translate(134217848 134217844)
  eval((keyboard-translate 134217848 134217844))
  eval-last-sexp-1(nil)
  eval-last-sexp(nil)
  call-interactively(eval-last-sexp nil nil)

Software:
- Windows XP SP3
- GNU Emacs 23.1.1 (i386-mingw-nt5.1.2600) of 2009-07-30 on SOFT-
MJASON

Any help?

Thanks.

Re: keyboard-translate working for Control but not for Alt

[harven]

Elena <egarrulo@gmail.com> writes:

> Hello,
>
> since I use a Dvorak keyboard, I'd like to swap C-x and C-t, as it is
> suggested here:
>
> http://www.emacswiki.org/emacs/DvorakKeyboard#toc4
>
> The proposed code performs its task:
>
> http://www.emacswiki.org/emacs/DvorakKeyboard#toc4
>
> (keyboard-translate ?\C-x ?\C-t)
> (keyboard-translate ?\C-t ?\C-x)
>
> Out of curiosity, I tried to swap M-x and M-t as well:
>
> (keyboard-translate ?\M-x ?\M-t)
> (keyboard-translate ?\M-t ?\M-x)

Use instead 

(global-set-key (kbd "M-x") 'execute-extended-command)
(global-set-key (kbd "M-t") 'transpose-words)

In contrast to C-x, M-x and M-t are not prefix keys.
So it is not necessary to use a keyboard-translation.

Re: keyboard-translate working for Control but not for Alt

[harven]

harven <harven@free.fr> writes:

> Elena <egarrulo@gmail.com> writes:
>>
>> Out of curiosity, I tried to swap M-x and M-t as well:
>>
>> (keyboard-translate ?\M-x ?\M-t)
>> (keyboard-translate ?\M-t ?\M-x)
>
> Use instead 
> (global-set-key (kbd "M-x") 'execute-extended-command)
> (global-set-key (kbd "M-t") 'transpose-words)>
>
> In contrast to C-x, M-x and M-t are not prefix keys.
> So it is not necessary to use a keyboard-translation.

 oups, use 

 (global-set-key (kbd "M-t") 'execute-extended-command)
 (global-set-key (kbd "M-x") 'transpose-words)

Re: keyboard-translate working for Control but not for Alt

[Xah Lee]

On Aug 4, 7:39 am, Elena <egarr...@gmail.com> wrote:
> Hello,
>
> since I use a Dvorak keyboard, I'd like to swap C-x and C-t, as it is
> suggested here:
>
> http://www.emacswiki.org/emacs/DvorakKeyboard#toc4
>
> The proposed code performs its task:
>
> http://www.emacswiki.org/emacs/DvorakKeyboard#toc4
>
> (keyboard-translate ?\C-x ?\C-t)
> (keyboard-translate ?\C-t ?\C-x)
>
> Out of curiosity, I tried to swap M-x and M-t as well:
>
> (keyboard-translate ?\M-x ?\M-t)
> (keyboard-translate ?\M-t ?\M-x)
>
> but evaluating each expression fail with (numbers are not the same):
>
> Debugger entered--Lisp error: (wrong-type-argument characterp
> 134217848)
>   keyboard-translate(134217848 134217844)
>   eval((keyboard-translate 134217848 134217844))
>   eval-last-sexp-1(nil)
>   eval-last-sexp(nil)
>   call-interactively(eval-last-sexp nil nil)
>
> Software:
> - Windows XP SP3
> - GNU Emacs 23.1.1 (i386-mingw-nt5.1.2600) of 2009-07-30 on SOFT-
> MJASON

just use this:

; change C-x prefix to C-8, so this frequently used prefix is easier
to type
(keyboard-translate ?\C-t ?\C-x)

this will make C-t do C-x, while C-x still remains valid.

Been using this in emacs since ~2001 on OS X, and recently Windows.
Been daily emacs users since 1998. Been dvorak typist since ~1992.

you might also be interested in a ergoemacs setup:
http://code.google.com/p/ergoemacs/

  Xah
∑ http://xahlee.org/

☄

Re: keyboard-translate working for Control but not for Alt

[Xah Lee]

On Aug 4, 7:39 am, Elena <egarr...@gmail.com> wrote:
> since I use a Dvorak keyboard,
> ...
> Out of curiosity, I tried to swap M-x and M-t as well:
>
> (keyboard-translate ?\M-x ?\M-t)
> (keyboard-translate ?\M-t ?\M-x)
>
> but evaluating each expression fail with (numbers are not the same):
>
> Debugger entered--Lisp error: (wrong-type-argument characterp
> 134217848)
>   keyboard-translate(134217848 134217844)
>   eval((keyboard-translate 134217848 134217844))
>   eval-last-sexp-1(nil)
>   eval-last-sexp(nil)
>   call-interactively(eval-last-sexp nil nil)
>
> Software:
> - Windows XP SP3
> - GNU Emacs 23.1.1 (i386-mingw-nt5.1.2600) of 2009-07-30 on SOFT-
> MJASON

Oops, didn't see your real question in my previous post.

Sorry don't have info on that except i have M-a as M-x. Not done as
keyboard-translate, but simply:
(global-set-key (kbd "M-a") 'execute-extended-command)

Note that the shortcuts M-x and C-x are rather logically in different
category.

M-x is not a prefix. It is just a shortcut to invoke command by name
(execute-extended-command). The “x” there is a mnemonic for eXtended.

While, C-x is a generic prefix to extend emacs's shortcuts system as
key combo sequences (meaning, you type a series of key combinations to
complete a shortcut. (e.g. C-x C-b, C-x 5 1, C-c C-f, C-h i))

The command shell-command (M-!) share more similarity to execute-
extended-command (M-x). So, it makes more sense to have these two
commands share similar shortcuts, either as a Shift and non-shift
version, or Ctrl vs Alt version. Using Shift for invoking similar/
reversal functions is conventional in modern apps. Using Ctrl vs Alt
for this purpose is traditional in emacs.

Emacs's M-x and C-x sharing the same letter “x” is rather a mistake i
think, since they dont have that much logical commonality. I think
this is also a major cause of confusion in emacs's shortcuts.

  Xah
∑ http://xahlee.org/

☄

Re: keyboard-translate working for Control but not for Alt

[Johan Bockgård]

Elena <egarrulo@gmail.com> writes:

> Out of curiosity, I tried to swap M-x and M-t as well:
>
> (keyboard-translate ?\M-x ?\M-t)
> (keyboard-translate ?\M-t ?\M-x)
>
> but evaluating each expression fail with (numbers are not the same):

keyboard-translate only works for characters.

You can use

  (define-key key-translation-map [?\M-x] [?\M-t])
  (define-key key-translation-map [?\M-t] [?\M-x])

Re: keyboard-translate working for Control but not for Alt

[Xah Lee]

no one has mentioned this yet...

if you do
(keyboard-translate ?\C-t ?\C-x)

one problem is that ALL your C-t keys becomes C-x.
So, for example, in outline-mode, C-c C-t runs hide-body.
But now, when you type that, you get C-c C-x instead.

A better solution is:
(global-set-key (kbd "C-t") ctl-x-map)

  Xah
∑ http://xahlee.org/

☄

On Aug 4, 7:39 am, Elena <egarr...@gmail.com> wrote:
> Hello,
>
> since I use a Dvorak keyboard, I'd like to swap C-x and C-t, as it is
> suggested here:
>
> http://www.emacswiki.org/emacs/DvorakKeyboard#toc4
>
> The proposed code performs its task:
>
> http://www.emacswiki.org/emacs/DvorakKeyboard#toc4
>
> (keyboard-translate ?\C-x ?\C-t)
> (keyboard-translate ?\C-t ?\C-x)
>
> Out of curiosity, I tried to swap M-x and M-t as well:
>
> (keyboard-translate ?\M-x ?\M-t)
> (keyboard-translate ?\M-t ?\M-x)
>
> but evaluating each expression fail with (numbers are not the same):
>
> Debugger entered--Lisp error: (wrong-type-argument characterp
> 134217848)
>   keyboard-translate(134217848 134217844)
>   eval((keyboard-translate 134217848 134217844))
>   eval-last-sexp-1(nil)
>   eval-last-sexp(nil)
>   call-interactively(eval-last-sexp nil nil)
>
> Software:
> - Windows XP SP3
> - GNU Emacs 23.1.1 (i386-mingw-nt5.1.2600) of 2009-07-30 on SOFT-
> MJASON
>
> Any help?
>
> Thanks.

Re: keyboard-translate working for Control but not for Alt

[Xah Lee]

Hi Elena,

; Make C-t acts like C-x, so that C-x is easier to type on dvorak
; (keyboard-translate ?\C-t ?\C-x)
; but this means  ALL your C-t keys becomes C-x.
; So, for example, in outline-mode, C-c C-t runs hide-body.
; But now, when you type that, you get C-c C-x instead.
; A better solution is:
; (global-set-key (kbd "C-t") ctl-x-map)

i've been using
(keyboard-translate ?\C-t ?\C-x)

for the past ~5 years or so. In pratice, i never noticed any problem,
prob because C-t hardly used as embedded combo in a key sequence, or,
when it is used, such commands are often advanced features or not
commonly needed.

in fact, the example C-c C-t in outline-mode i only noticed yesterday,
as i was browsing emacs 23's News file which is in outline-mode. Other
than this, i've not used outline-mode.

I switched to

(global-set-key (kbd "C-t") ctl-x-map)

yesterday, and noticed that it has problems too.
I often do “C-x r l” to get my bookmark list.
So, this means i've been pressing “C-t r l” in the past ~5 years.
But if you are in dired, that no longer works with the new setup. “C-t
r” invokes image-dired-delete-tag.

  Xah
∑ http://xahlee.org/

☄

Re: keyboard-translate working for Control but not for Alt

[Xah Lee]

additional info i just found out ....

this method
(global-set-key (kbd "C-t") ctl-x-map)

turns out to have other problems as well, when cua-mode is on.

with cua-mode on, and when you have a text selection, you need to
press C-x twice quickly to invoke the traditional C-x role. (at least
as one of the method, anyway) So, that means you press C-t twice
quickly.

One example i'll need to do this many times per day is “C-t r
t” (string-rectangle) and “C-t r k” for kill-rectangle, where you do
need a selection on first.

With
 (keyboard-translate ?\C-t ?\C-x)
i used in the past, i didn't encounter problem.

i started to use cua-mode since last year, not because i need it, but
more for testing compatibilty with my ergoemacs setup.

So, with cua-mode on, and with
(global-set-key (kbd "C-t") ctl-x-map),
now “C-t r t” no longer works... it seems to be calling transpose-
lines...

in short, all these extra personalized fixes creates a lot
complications. If emacs dev made cua mode's intended purpose by
default, which as a side effect most of these complications will be
gone, as compared to indivdualistic customization that tried to
achieves the same thing. This is because when emacs has something out
of the box, the implementation is more throughly examed at some lower
level, i.e. creating much more robustness than individual hacks. If
emacs dev decided that Ctrl + XCV should be the shortcut for emacs,
the proper solution is probably not even using cua-mode, since itself
is a quite complex hack.

  Xah
∑ http://xahlee.org/

☄

On Aug 5, 3:37 pm, Xah Lee <xah...@gmail.com> wrote:
> Hi Elena,
>
> ; Make C-t acts like C-x, so that C-x is easier to type on dvorak
> ; (keyboard-translate ?\C-t ?\C-x)
> ; but this means  ALL your C-t keys becomes C-x.
> ; So, for example, in outline-mode, C-c C-t runs hide-body.
> ; But now, when you type that, you get C-c C-x instead.
> ; A better solution is:
> ; (global-set-key (kbd "C-t") ctl-x-map)
>
> i've been using
> (keyboard-translate ?\C-t ?\C-x)
>
> for the past ~5 years or so. In pratice, i never noticed any problem,
> prob because C-t hardly used as embedded combo in a key sequence, or,
> when it is used, such commands are often advanced features or not
> commonly needed.
>
> in fact, the example C-c C-t in outline-mode i only noticed yesterday,
> as i was browsing emacs 23's News file which is in outline-mode. Other
> than this, i've not used outline-mode.
>
> I switched to
>
> (global-set-key (kbd "C-t") ctl-x-map)
>
> yesterday, and noticed that it has problems too.
> I often do “C-x r l” to get my bookmark list.
> So, this means i've been pressing “C-t r l” in the past ~5 years.
> But if you are in dired, that no longer works with the new setup. “C-t
> r” invokes image-dired-delete-tag.
>
>   Xah
> ∑http://xahlee.org/
>
> ☄

Re: keyboard-translate working for Control but not for Alt

[Xah Lee]

i cleaned up my writing a bit, hopefully it's more readible as a
summary of the issues and solutions.

• Emacs, Dvorak, C-x
  http://xahlee.org/emacs/emacs_dvorak_C-x.html

  Xah
∑ http://xahlee.org/

☄

Re: keyboard-translate working for Control but not for Alt

[B. T. Raven]

Elena and Xah:

A conservative approach to Emacs ergonomics with a Dvorak keyboard on 
w32  is to pull the keycaps, including the bottom row and use Keytweak 
to reassign the mod keys:

super alt ctl spacebar ctl alt super hyper

this along with something like the following added to .emacs allows for 
minimum secondguessing of the Emacs key-binding philosophy as it has 
evolved over the last 20 years:

;; Single char cursor movement on Dvorak layout
(global-set-key [(meta h)] 'backward-char)
(global-set-key [(meta n)] 'forward-char)
(global-set-key [(meta c)] 'previous-line)
(global-set-key [(meta t)] 'next-line)
(global-set-key [(meta H)] 'mark-paragraph)
(global-set-key [(meta N)] 'next-buffer)
(global-set-key [(meta C)] 'capitalize-word)
(global-set-key [(meta T)] 'transpose-words)

;;substitute for stolen metakeychords ad libitum
(global-set-key [(control n)] 'next-line-mark)
(global-set-key [(control p)] 'center-paragraph)
(global-set-key [(control b)] 'beginning-of-line-mark)
(global-set-key [(control f)] 'find-function-at-point)
(global-set-key [(shift control f)] 'find-variable-at-point)

If all mod keys could be placed for easy access from the home keys and 
if there were some way to make the simultaneous pressing of 3 mod keys 
sticky (i.e allowing further mod keys to be stacked on top of them) then 
one could touch type things like shift control alt super hyper meta a



Xah Lee wrote:
> additional info i just found out ....
> 
> this method
> (global-set-key (kbd "C-t") ctl-x-map)
> 
> turns out to have other problems as well, when cua-mode is on.
> 
> with cua-mode on, and when you have a text selection, you need to
> press C-x twice quickly to invoke the traditional C-x role. (at least
> as one of the method, anyway) So, that means you press C-t twice
> quickly.
> 
> One example i'll need to do this many times per day is “C-t r
> t” (string-rectangle) and “C-t r k” for kill-rectangle, where you do
> need a selection on first.
> 
> With
>  (keyboard-translate ?\C-t ?\C-x)
> i used in the past, i didn't encounter problem.
> 
> i started to use cua-mode since last year, not because i need it, but
> more for testing compatibilty with my ergoemacs setup.
> 
> So, with cua-mode on, and with
> (global-set-key (kbd "C-t") ctl-x-map),
> now “C-t r t” no longer works... it seems to be calling transpose-
> lines...
> 
> in short, all these extra personalized fixes creates a lot
> complications. If emacs dev made cua mode's intended purpose by
> default, which as a side effect most of these complications will be
> gone, as compared to indivdualistic customization that tried to
> achieves the same thing. This is because when emacs has something out
> of the box, the implementation is more throughly examed at some lower
> level, i.e. creating much more robustness than individual hacks. If
> emacs dev decided that Ctrl + XCV should be the shortcut for emacs,
> the proper solution is probably not even using cua-mode, since itself
> is a quite complex hack.
> 
>   Xah
> ∑ http://xahlee.org/
> 
> ☄
> 
> On Aug 5, 3:37 pm, Xah Lee <xah...@gmail.com> wrote:
>> Hi Elena,
>>
>> ; Make C-t acts like C-x, so that C-x is easier to type on dvorak
>> ; (keyboard-translate ?\C-t ?\C-x)
>> ; but this means  ALL your C-t keys becomes C-x.
>> ; So, for example, in outline-mode, C-c C-t runs hide-body.
>> ; But now, when you type that, you get C-c C-x instead.
>> ; A better solution is:
>> ; (global-set-key (kbd "C-t") ctl-x-map)
>>
>> i've been using
>> (keyboard-translate ?\C-t ?\C-x)
>>
>> for the past ~5 years or so. In pratice, i never noticed any problem,
>> prob because C-t hardly used as embedded combo in a key sequence, or,
>> when it is used, such commands are often advanced features or not
>> commonly needed.
>>
>> in fact, the example C-c C-t in outline-mode i only noticed yesterday,
>> as i was browsing emacs 23's News file which is in outline-mode. Other
>> than this, i've not used outline-mode.
>>
>> I switched to
>>
>> (global-set-key (kbd "C-t") ctl-x-map)
>>
>> yesterday, and noticed that it has problems too.
>> I often do “C-x r l” to get my bookmark list.
>> So, this means i've been pressing “C-t r l” in the past ~5 years.
>> But if you are in dired, that no longer works with the new setup. “C-t
>> r” invokes image-dired-delete-tag.
>>
>>   Xah
>> ∑http://xahlee.org/
>>
>> ☄

Re: keyboard-translate working for Control but not for Alt

[Elena]

On 6 Ago, 18:34, "B. T. Raven" <ni...@nihilo.net> wrote:
> A conservative approach to Emacs ergonomics with a Dvorak keyboard on
> w32  is to pull the keycaps, including the bottom row and use Keytweak
> to reassign the mod keys:
>
> super alt ctl spacebar ctl alt super hyper

What do you use the Hyper key for?

> If all mod keys could be placed for easy access from the home keys and
> if there were some way to make the simultaneous pressing of 3 mod keys
> sticky (i.e allowing further mod keys to be stacked on top of them) then
> one could touch type things like shift control alt super hyper meta a

Yes, Windows allows for sticky modifier: you can set that into
Accessibility options (Control Panel).

Thanks.

Re: keyboard-translate working for Control but not for Alt

[Elena]

On 6 Ago, 23:04, Elena <egarr...@gmail.com> wrote:
> On 6 Ago, 18:34, "B. T. Raven" <ni...@nihilo.net> wrote:
>
> > A conservative approach to Emacs ergonomics with a Dvorak keyboard on
> > w32  is to pull the keycaps, including the bottom row and use Keytweak
> > to reassign the mod keys:
>
> > super alt ctl spacebar ctl alt super hyper
>
> What do you use the Hyper key for?

Wait a moment! I don't think that Hyper is available on Windows... is
it?

Thanks

Re: keyboard-translate working for Control but not for Alt

[B. T. Raven]

Elena wrote:
> On 6 Ago, 23:04, Elena <egarr...@gmail.com> wrote:
>> On 6 Ago, 18:34, "B. T. Raven" <ni...@nihilo.net> wrote:
>>
>>> A conservative approach to Emacs ergonomics with a Dvorak keyboard on
>>> w32  is to pull the keycaps, including the bottom row and use Keytweak
>>> to reassign the mod keys:
>>> super alt ctl spacebar ctl alt super hyper
>> What do you use the Hyper key for?
> 
> Wait a moment! I don't think that Hyper is available on Windows... is
> it?
> 
> Thanks

You also need this in your .emacs:

(setq w32-pass-lwindow-to-system nil
       w32-pass-rwindow-to-system nil
       w32-pass-apps-to-system    nil
       w32-lwindow-modifier       'super   ;; Left Windows
       w32-rwindow-modifier       'super   ;; Right Windows
       w32-apps-modifier          'hyper)  ;; App-Menu (key to right of 
Right Windows)

I discovered that I don't use hyper for anything. But I *could*. Btw, 
even with this setup, a lot of mswin key combos leak through, making 
them unavailable to Emacs, but if you look at them a little cross-eyed 
they can be seen as features rather than bugs. E.g. repeated Alt-tab 
cycles through loaded windows apps whether you want to or not.


Ed

Re: keyboard-translate working for Control but not for Alt

[B. T. Raven]

Elena wrote:
> On 6 Ago, 18:34, "B. T. Raven" <ni...@nihilo.net> wrote:
>> A conservative approach to Emacs ergonomics with a Dvorak keyboard on
>> w32  is to pull the keycaps, including the bottom row and use Keytweak
>> to reassign the mod keys:
>>
>> super alt ctl spacebar ctl alt super hyper
> 
> What do you use the Hyper key for?
> 
>> If all mod keys could be placed for easy access from the home keys and
>> if there were some way to make the simultaneous pressing of 3 mod keys
>> sticky (i.e allowing further mod keys to be stacked on top of them) then
>> one could touch type things like shift control alt super hyper meta a
> 
> Yes, Windows allows for sticky modifier: you can set that into
> Accessibility options (Control Panel).
> 
> Thanks.

Yeah, I knew about that but I wanted them to be sticky dynamically, only 
if 3 were pressed simultaneously. Three plus an alphanumeric is about 
the limit for comfortable touch typing. As soon as the keychord is 
interpreted by Emacs they should become unsticky again.

Re: keyboard-translate working for Control but not for Alt

[Elena]

On 4 Ago, 18:36, bojohan+n...@dd.chalmers.se (Johan Bockgård) wrote:
> Elena <egarr...@gmail.com> writes:
> > Out of curiosity, I tried to swap M-x and M-t as well:
>
> > (keyboard-translate ?\M-x ?\M-t)
> > (keyboard-translate ?\M-t ?\M-x)
>
> > but evaluating each expression fail with (numbers are not the same):
>
> keyboard-translate only works for characters.
>
> You can use
>
>   (define-key key-translation-map [?\M-x] [?\M-t])
>   (define-key key-translation-map [?\M-t] [?\M-x])

That does not work. No errors when evaluating the expressions, but
then both M-x and M-t don't work anymore (they just output 't' and
'x').

Thanks for your answer.

Re: keyboard-translate working for Control but not for Alt

[Bernardo]

>>
>> You can use
>>
>>   (define-key key-translation-map [?\M-x] [?\M-t])
>>   (define-key key-translation-map [?\M-t] [?\M-x])
> 
> That does not work. No errors when evaluating the expressions, but
> then both M-x and M-t don't work anymore (they just output 't' and
> 'x').


don't know, both the above and the following format work here (GNU/Linux):
(define-key key-translation-map [(meta x)] [(meta t)])

not sure if it's any different under the hood but you may also want to 
try this variation:
(keyboard-translate ?\M-x ?\M-t)

Re: keyboard-translate working for Control but not for Alt

[Elena]

On 6 Ago, 11:01, Bernardo <bernardo.ba...@pobox.com> wrote:
> >> You can use
>
> >>   (define-key key-translation-map [?\M-x] [?\M-t])
> >>   (define-key key-translation-map [?\M-t] [?\M-x])
>
> > That does not work. No errors when evaluating the expressions, but
> > then both M-x and M-t don't work anymore (they just output 't' and
> > 'x').
>
> don't know, both the above and the following format work here (GNU/Linux):
> (define-key key-translation-map [(meta x)] [(meta t)])
>

I apologize for not testing it properly... it works under "emacs -q".
Something is broken in my customization.

Thanks for pointing that out.

> not sure if it's any different under the hood but you may also want to
> try this variation:
> (keyboard-translate ?\M-x ?\M-t)

This raises an error with "emacs -q" too.

Thanks for your reply.