troubleshoot C-M-/ key binding (dabbrev-completion)

troubleshoot C-M-/ key binding (dabbrev-completion)

[Michael V. Antosha]

Hello,

Key binding C-M-/ (dabbrev-completion) does not seem to work for me.
Could someone point me in a right direction to find the cause of this behavior?


Details:

1. dabbrev-completion works fine if used explicitly (M-x dabbrev-completion).

2. M-/ (dabbrev-expand) and C-/ (undo) both work as expected.

3. Using C-M-/ (i.e. pressing Ctrl+Alt+SLASH) seems to have the same
effect as M-/ (dabbrev-expand).

4. Other C-M-... combinations seem to work fine.

For example, C-M-\ (indent-region), C-\ (toggle-input-method) and M-\
(delete-horizontal-space) all work as expected, each doing its own
function.

5. From the output of describe-bindings (C-h b):

    M-/             dabbrev-expand
    C-M-/           dabbrev-completion
    C-/             undo

    C-\             toggle-input-method
    M-\             delete-horizontal-space
    C-M-\           indent-region


Any hints?
I want to find out what is wrong with my setup
(and how to troubleshoot -- step by step -- similar issues with Emacs).

Thank you in advance!

Best regards,
Michael


-- 
Michael V. Antosha

Re: troubleshoot C-M-/ key binding (dabbrev-completion)

[Martin Marshall]

"Michael V. Antosha" <michael.antosha@gmail.com> writes:

> 3. Using C-M-/ (i.e. pressing Ctrl+Alt+SLASH) seems to have the same
> effect as M-/ (dabbrev-expand).

What effect is that?

Are you sure there are multiple completion candidates in the buffer?  If
there's only one candidate in the buffer, `dabbrev-completion' will
immediately complete the word without bringing up a "*Completions*"
buffer.  That would look like the same thing `dabbrev-expand' does.

You can check which key Emacs thinks you pressed and which command it
called by invoking `view-lossage' (bound to "C-h l").

-- 
Best regards,
Martin Marshall

Re: troubleshoot C-M-/ key binding (dabbrev-completion)

[Michael V. Antosha]

On 1/21/24, Martin Marshall <law@martinmarshall.com> wrote:
> "Michael V. Antosha" <michael.antosha@gmail.com> writes:
>
>> 3. Using C-M-/ (i.e. pressing Ctrl+Alt+SLASH) seems to have the same
>> effect as M-/ (dabbrev-expand).
>
> What effect is that?

I prepared a description of a test procedure to specify the effect,
but I chose to skip sending it to keep my post shorter and because it
seems unnecessary now (explained below).


> Are you sure there are multiple completion candidates in the buffer?

Yes.


> If there's only one candidate in the buffer, `dabbrev-completion' will
> immediately complete the word without bringing up a "*Completions*"
> buffer.  That would look like the same thing `dabbrev-expand' does.

I understand differences and similarities between the two commands, I
first tested their behavior without key bindings, through M-x
dabbrev-*


> You can check which key Emacs thinks you pressed and which command it
> called by invoking `view-lossage' (bound to "C-h l").

Martin, thank you very much!
That's really useful.


What I'm seeing now proves my original suspicion:
 - when I press "Alt+SLASH" emacs interprets it as ESC / [dabbrev-expand]
 - when I press "Ctrl+Alt+SLASH" emacs interprets it, as ESC / [dabbrev-expand]

 ESC / [dabbrev-expand]
 C-h l [view-lossage]

What could I do further to find out why these different combinations
are understood by emacs as ESC / sequence?


For comparison, Ctrl+Alt+BACKSLASH works fine.
Below are keys/commands from C-h l after pressing Ctr+BACKSLASH, then
Alt+BACKSLASH, then Ctrl+Alt+BACKSLASH:

 C-\ [toggle-input-method]
 ...
 ESC \ [delete-horizontal-space]
 ...
 ESC C-\ [indent-region]
 C-h l [view-lossage]


I also tried to emulate C-M-/ by pressing ESC and then pressing Ctrl+SLASH,
the result is quite unexpected:

 ESC C-_ [nil]
 C-h l [view-lossage]


-- 
Michael V. Antosha

Re: troubleshoot C-M-/ key binding (dabbrev-completion)

[Pierre Rouleau]

Hi Michael,

Are you using Emacs in terminal mode?
If you are, it's quite possible that any key sequence that does not map to
ASCII Control characters will not work.
Note that C-\ is a valid ASCII Control code (hex byte value 0x5c).
And C-\ is not a valid ASCII control code.
Therefore Emacs, in terminal mode will not be able to recognize C-M-\

In graphics mode, Emacs will be able to identify C-M-\ as it will decode
the key sequence from the keyboard
directly and won't have to use a terminal emulator standard.

To see if a key sequence is recognized, type C-h k  followed by the key
sequence you're looking for.
Type C-h k C-M-\   to see the mapping of that key sequence and whether it
is recognized.

Hope this helps,

/Pierre



On Sun, Jan 21, 2024 at 5:00 PM Michael V. Antosha <
michael.antosha@gmail.com> wrote:

> On 1/21/24, Martin Marshall <law@martinmarshall.com> wrote:
> > "Michael V. Antosha" <michael.antosha@gmail.com> writes:
> >
> >> 3. Using C-M-/ (i.e. pressing Ctrl+Alt+SLASH) seems to have the same
> >> effect as M-/ (dabbrev-expand).
> >
> > What effect is that?
>
> I prepared a description of a test procedure to specify the effect,
> but I chose to skip sending it to keep my post shorter and because it
> seems unnecessary now (explained below).
>
>
> > Are you sure there are multiple completion candidates in the buffer?
>
> Yes.
>
>
> > If there's only one candidate in the buffer, `dabbrev-completion' will
> > immediately complete the word without bringing up a "*Completions*"
> > buffer.  That would look like the same thing `dabbrev-expand' does.
>
> I understand differences and similarities between the two commands, I
> first tested their behavior without key bindings, through M-x
> dabbrev-*
>
>
> > You can check which key Emacs thinks you pressed and which command it
> > called by invoking `view-lossage' (bound to "C-h l").
>
> Martin, thank you very much!
> That's really useful.
>
>
> What I'm seeing now proves my original suspicion:
>  - when I press "Alt+SLASH" emacs interprets it as ESC / [dabbrev-expand]
>  - when I press "Ctrl+Alt+SLASH" emacs interprets it, as ESC /
> [dabbrev-expand]
>
>  ESC / [dabbrev-expand]
>  C-h l [view-lossage]
>
> What could I do further to find out why these different combinations
> are understood by emacs as ESC / sequence?
>
>
> For comparison, Ctrl+Alt+BACKSLASH works fine.
> Below are keys/commands from C-h l after pressing Ctr+BACKSLASH, then
> Alt+BACKSLASH, then Ctrl+Alt+BACKSLASH:
>
>  C-\ [toggle-input-method]
>  ...
>  ESC \ [delete-horizontal-space]
>  ...
>  ESC C-\ [indent-region]
>  C-h l [view-lossage]
>
>
> I also tried to emulate C-M-/ by pressing ESC and then pressing Ctrl+SLASH,
> the result is quite unexpected:
>
>  ESC C-_ [nil]
>  C-h l [view-lossage]
>
>
> --
> Michael V. Antosha
>
>

-- 
/Pierre