On 2023-04-28 21:50, R.Wieser wrote:
> Carlos,
>
>> Without knowing really how the keyboard works in Linux (I knew it in depth
>> for MsDOS), the keys like "control" are not normal keys, they act outside
>> the matrix, changing the bit value of some word.
>
> :-) If it would have been DOS than I would just have written a TSR to
> translate keys to whatever I wanted.
>
> My problem is similar to yours : I'm just too much of a newbie in regard to
> Linux., and simply do not know squat about stuff like this.

I have been using Linux since 1997, but there are many things that I
don't know details. Keyboard is one of them. I've never understood why
ctrl-arrows did not, and still does not, work in several text editors.
Or they work in a terminal but not in a console. It is beyond my
understanding.

(try using 'mc' in console: at least in openSUSE, the function keys
change positions and there is actual danger of deleting files by accident).

Long ago I tried to understand it. And I have forgotten a lot of it.

Apparently, Linux inherits this from Unix, and it tries to emulate old
hardware terminal machines that followed certain protocols. One was the
VT100. But then Linux has its own, named "linux", IIRC.

For example, in the XFCE terminal, I see this:

cer@Telcontar:~> set | grep TERM
COLORTERM=truecolor
TERM=xterm-256color
cer@Telcontar:~>

and on my console I get:

cer@Telcontar:~> cat p
COLORTERM=1
TERM=linux

So it behaves differently down there.

Ah, for positioning coloured letters, and box drawing, in text terminal
and consoles, doing menus in text, or reading the mouse in text, things
get interesting. You have to use "CURSES", but there are several
implementations, and it is not the only method.

Something that MsDOS did in the 80's and Linux almost still doesn't, but
for a few exceptions.

>> But you can compare 'a' with '^a' to find out.
>
> the vt100(?) sequence itself wasn't the problem - I already had that and
> confirmed it using "showkey -a" - it was where to place it.
>
> Just before you Chris posted information on how to do it for SuSe, and it
> gave me enough information to be able to find the way to do it on RPi
> bullseye.

Ah. I'll seek for it. [...] Ah, the link he posted.

The information is obsolete for openSUSE, the file
/etc/sysconfig/keyboard does not contain "KEYTABLE" now. I have:

YAST_KEYBOARD="spanish,pc104"

which defines my keyboard in console. I have not found a file named
pc104, except a .h file in the kernel sources.

--
Cheers, Carlos.

On 28/04/2023 20:39, R.Wieser wrote:
> Chris,
>
>> Does this help?
>> http://www.linuxproblem.org/art_17.html
>
> It does, very even.
>
> The only difference is that in Rpi Bullseye the file is
>
> /etc/default/keyboard
>
> , and the line that needs to be added is
>
> KMAP=/usr/local/mykeymap
>
> (found the "kmap" entry in "man 5 keyboard")
>
> Thanks.
>
> Regards,
> Rudy Wieser
>
>

Yes. It was meant to help. Not to do it for you <g>.
You could also have put the command:
loadkeys /usr/local/mykeymap
into rc.local (for those of us without /etc/default/keyboard)

--
Chris Elvidge
England

Carlos,

> I've never understood why ctrl-arrows did not, and still does not, work
> in several text editors.

As far as I know a lot of "extra" keys only "quite recently" where given a
fixed keyscan code. Way-back-when I wrote a program which used keyboad
input and worked well, but whent to shite when I had to replace the
keyboard.

> Or they work in a terminal but not in a console. It is beyond my
> understanding.

You said it yourself : linux tried to make sure it wold work on the
origional, old hardware - keyboards which did not have those keys - and most
writers of editors who where more focussed on recent hardware.

> Ah, for positioning coloured letters, and box drawing, in text terminal
> and consoles, doing menus in text, or reading the mouse in text, things
> get interesting. You have to use "CURSES", but there are several
> implementations, and it is not the only method.

I would have liked to use a (hopefully standard) library like that. Alas,
bullseye lite did not come with it, so now I'm trying my hand at drawing
text-mode boxes the good-old way, using those vt100(?) sequences.

Not that I have a probelm with doing low-level stuff like that mind you.
Starting at the bottom will create understanding of how it all works as well
as an appriciation to how much work the diffent libraries do for me.

> The information is obsolete for openSUSE, the file /etc/sysconfig/keyboard
> does not contain "KEYTABLE" now. I have:
>
> YAST_KEYBOARD="spanish,pc104"

Its quite possible that the KEYTABLE entry is only needed if you want to
override the default, build-in one. At least, in my case there was no
KMAP entry, and it had to be added (pointing to the custom keymap)

Regards,
Rudy Wieser

Chris,

> Yes. It was meant to help. Not to do it for you <g>.

What ? You post something to only /help/ and not spoon-feed it to me ?
I'm appalled sir, appalled I tell you. :-D

To be honest, when I reply to something thats most always what I try to do
too : help. Spoon-feeding creates people who will expect you to keep doing
that, and will sometimes agressivily respond when you tell them they have to
put some work in it themselves too. :-\

> You could also have put the command:
> loadkeys /usr/local/mykeymap
> into rc.local (for those of us without /etc/default/keyboard)

Question : Do you know of a technical reason why I should load the full
exported, rather big keymap instead of just the "wordnav.keys" one (containg
just the changes) ?

Full disclosure : I was already thinking of doing the latter, but first
needed to figure out where to do it. Something just "spoon feeded" to me.
:-)

Regards,
Rudy Wieser

P.s.
Just found out that "showkeys -k" will display the keycode for a key,
meaning that I've got all the info to needed to create other translations
too.

On 29/04/2023 07:16, R.Wieser wrote:
> Chris,
>
>> Yes. It was meant to help. Not to do it for you <g>.
>
> What ? You post something to only /help/ and not spoon-feed it to me ?
> I'm appalled sir, appalled I tell you. :-D
>
> To be honest, when I reply to something thats most always what I try to do
> too : help. Spoon-feeding creates people who will expect you to keep doing
> that, and will sometimes agressivily respond when you tell them they have to
> put some work in it themselves too. :-\
>
>> You could also have put the command:
>> loadkeys /usr/local/mykeymap
>> into rc.local (for those of us without /etc/default/keyboard)
>
> Question : Do you know of a technical reason why I should load the full
> exported, rather big keymap instead of just the "wordnav.keys" one (containg
> just the changes) ?

You're loading the default full keymap and then the wordnav.keys
OR
You're loading the exported full keymap - which overwrites the default.

Either way, whatever seems best to you.
Remember if you modify a file that is provided by the OS, it _may_ be
overwritten by an upgrade.

>
> Full disclosure : I was already thinking of doing the latter, but first
> needed to figure out where to do it. Something just "spoon feeded" to me.
> :-)
>
> Regards,
> Rudy Wieser
>
> P.s.
> Just found out that "showkeys -k" will display the keycode for a key,
> meaning that I've got all the info to needed to create other translations
> too.
>

Good

>

--
Chris Elvidge
England

On 2023-04-29 07:37, R.Wieser wrote:
> Carlos,
>
>> I've never understood why ctrl-arrows did not, and still does not, work
>> in several text editors.
>
> As far as I know a lot of "extra" keys only "quite recently" where given a
> fixed keyscan code. Way-back-when I wrote a program which used keyboad
> input and worked well, but whent to shite when I had to replace the
> keyboard.
>
>> Or they work in a terminal but not in a console. It is beyond my
>> understanding.
>
> You said it yourself : linux tried to make sure it wold work on the
> origional, old hardware - keyboards which did not have those keys - and most
> writers of editors who where more focussed on recent hardware.

Or just IBM PC clones and offspring.

>> Ah, for positioning coloured letters, and box drawing, in text terminal
>> and consoles, doing menus in text, or reading the mouse in text, things
>> get interesting. You have to use "CURSES", but there are several
>> implementations, and it is not the only method.
>
> I would have liked to use a (hopefully standard) library like that. Alas,
> bullseye lite did not come with it, so now I'm trying my hand at drawing
> text-mode boxes the good-old way, using those vt100(?) sequences.

It doesn't have any curses?

I have all these:

# rpm -qa | grep -i curses
python-curses-2.7.18-150000.48.1.x86_64
libyui-ncurses-pkg16-4.3.3-150400.1.8.x86_64
python3-curses-3.6.15-150300.10.45.1.x86_64
libncurses5-6.1-150000.5.12.1.x86_64
libyui-ncurses15-4.1.5-150300.3.8.7.x86_64
ncurses-utils-6.1-150000.5.12.1.x86_64
libyui-ncurses11-2.54.5-1.36.x86_64
libncurses6-6.1-150000.5.12.1.x86_64
ncurses-devel-6.1-150000.5.12.1.x86_64
libyui-ncurses16-4.3.3-150400.1.5.x86_64
#

I think you need libncurses (new curses, I think)

> Not that I have a probelm with doing low-level stuff like that mind you.
> Starting at the bottom will create understanding of how it all works as well
> as an appriciation to how much work the diffent libraries do for me.
>
>> The information is obsolete for openSUSE, the file /etc/sysconfig/keyboard
>> does not contain "KEYTABLE" now. I have:
>>
>> YAST_KEYBOARD="spanish,pc104"
>
> Its quite possible that the KEYTABLE entry is only needed if you want to
> override the default, build-in one. At least, in my case there was no
> KMAP entry, and it had to be added (pointing to the custom keymap)

I suspect the entry was there in the past, got removed.

--
Cheers, Carlos.

Chris,

>> Question : Do you know of a technical reason why I should load the full
>> exported, rather big keymap instead of just the "wordnav.keys" one
>> (containg just the changes) ?
....
> Either way, whatever seems best to you.

IOW, no technical reason not to use the "small change" method. Thanks.

> Remember if you modify a file that is provided by the OS, it _may_ be
> overwritten by an upgrade.

I know. There are a few reasons why I never upgrade, and that is one of
them. I do happen to download and install new versions though.

Regards,
Rudy Wieser

Carlos,

> It doesn't have any curses?

A system-wide search does turn up some hits, but none in /usr/include .

> # rpm -qa | grep -i curses

Currently I have no idea how check for or apply such "ready to install"
packages (if that is what they are). I'm a bit of a newbie I'm afraid. :-\

Regards,
Rudy Wieser

On 2023-04-29 19:26, R.Wieser wrote:
> Carlos,
>
>> It doesn't have any curses?
>
> A system-wide search does turn up some hits, but none in /usr/include .

Ah, no, for includes, in openSUSE parlance, you need install some -devel
package. I mean, something named like "libncurses5-devel".

>> # rpm -qa | grep -i curses
>
> Currently I have no idea how check for or apply such "ready to install"
> packages (if that is what they are). I'm a bit of a newbie I'm afraid. :-\

In raspberian, I have no idea.

--
Cheers, Carlos.

On 29/04/2023 19:06, Carlos E.R. wrote:
> On 2023-04-29 19:26, R.Wieser wrote:
>> Carlos,
>>
>>> It doesn't have any curses?
>>
>> A system-wide search does turn up some hits, but none in /usr/include .
>
> Ah, no, for includes, in openSUSE parlance, you need install some -devel
> package. I mean, something named like "libncurses5-devel".
>
>
>>> # rpm -qa | grep -i curses
>>
>> Currently I have no idea how check for or apply such "ready to install"
>> packages (if that is what they are). I'm a bit of a newbie I'm
>> afraid. :-\
>
> In raspberian, I have no idea.
>

On mine: `apt search ncurses` gives:

ncurses-base/stable,now 6.2+20201114-2 all [installed]
basic terminal type definitions

ncurses-bin/stable,now 6.2+20201114-2 armhf [installed]
terminal-related programs and man pages

ncurses-doc/stable 6.2+20201114-2 all
developer's guide and documentation for ncurses

ncurses-examples/stable 6.2+20201114-2 armhf
test programs and examples for ncurses

ncurses-hexedit/stable 0.9.7+orig-7.1 armhf
Edit files/disks in hex, ASCII and EBCDIC

ncurses-term/stable,now 6.2+20201114-2 all [installed,automatic]
additional terminal type definitions

So `apt install ncurses*` will do it

--
Chris Elvidge
England

Chris,

> On mine: `apt search ncurses` gives:
<snip>

On mine I get about 170 entries, with quite a few not even having "ncurses"
in the filename :

Filtering on "ncurses" in the filename shrinks that to 37. Selecting only
files starting with "ncurses" brings that down to 6, with 3 not already
installed :

ncurses-doc/stable 6.2+20201114-2 all
ncurses-examples/stable 6.2+20201114-2 armhf
ncurses-hexedit/stable 0.9.7+orig-7.1 armhf

But none of the above have "dev(el(op)) in their filenames ... :-|

When I look thru the 37 and search for "dev" I get these :

libcunit1-ncurses-dev/stable 2.1-3-dfsg-2.3 armhf
libcurses-ocaml-dev/stable 1.0.6-2+b1 armhf
libncurses-dev/stable 6.2+20201114-2 armhf
libncurses5-dev/stable 6.2+20201114-2 armhf
libncursesada9-dev/stable 6.2.20200212-4 armhf
libncursesw5-dev/stable 6.2+20201114-2 armhf
libnotcurses-dev/stable 2.2.3+dfsg.1-1 armhf
librust-ncurses-dev/stable 5.99.0-3 armhf

If I would have to guess (yuck) and pick one I would go for the third,
libncurses-dev

I take it that installing the "docs" and "examples" would be a good idea
too, unless thats already part of the "dev" install or a dependancy.

A good guess or am I way off base ?

Regards,
Rudy Wieser

On 2023-04-30 07:20, R.Wieser wrote:
> Chris,
>
>> On mine: `apt search ncurses` gives:
> <snip>
>
> On mine I get about 170 entries, with quite a few not even having "ncurses"
> in the filename :
>
> Filtering on "ncurses" in the filename shrinks that to 37. Selecting only
> files starting with "ncurses" brings that down to 6, with 3 not already
> installed :
>
> ncurses-doc/stable 6.2+20201114-2 all
> ncurses-examples/stable 6.2+20201114-2 armhf
> ncurses-hexedit/stable 0.9.7+orig-7.1 armhf
>
> But none of the above have "dev(el(op)) in their filenames ... :-|
>
> When I look thru the 37 and search for "dev" I get these :
>
> libcunit1-ncurses-dev/stable 2.1-3-dfsg-2.3 armhf
> libcurses-ocaml-dev/stable 1.0.6-2+b1 armhf
> libncurses-dev/stable 6.2+20201114-2 armhf
> libncurses5-dev/stable 6.2+20201114-2 armhf
> libncursesada9-dev/stable 6.2.20200212-4 armhf
> libncursesw5-dev/stable 6.2+20201114-2 armhf
> libnotcurses-dev/stable 2.2.3+dfsg.1-1 armhf
> librust-ncurses-dev/stable 5.99.0-3 armhf

My outsider guess is that "dev" means here something different than it
does for openSUSE, when I see the wording "dev/stable". They are
probably experimental packages.

You need help from somebody familiar with that distribution who knows
what packages contain the header files, how to find them.

--
Cheers, Carlos.

On 4/30/2023 4:58 AM, Carlos E.R. wrote:
> On 2023-04-30 07:20, R.Wieser wrote:
>> Chris,
>>
>>> On mine: `apt search ncurses` gives:
>> <snip>
>>
>> On mine I get about 170 entries, with quite a few not even having "ncurses"
>> in the filename :
>>
>> Filtering on "ncurses" in the filename shrinks that to 37.   Selecting only
>> files starting with "ncurses" brings that down to 6, with 3 not already
>> installed :
>>
>> ncurses-doc/stable 6.2+20201114-2 all
>> ncurses-examples/stable 6.2+20201114-2 armhf
>> ncurses-hexedit/stable 0.9.7+orig-7.1 armhf
>>
>> But none of the above have "dev(el(op)) in their filenames ... :-|
>>
>> When I look thru the 37 and search for "dev" I get these :
>>
>> libcunit1-ncurses-dev/stable 2.1-3-dfsg-2.3 armhf
>> libcurses-ocaml-dev/stable 1.0.6-2+b1 armhf
>> libncurses-dev/stable 6.2+20201114-2 armhf
>> libncurses5-dev/stable 6.2+20201114-2 armhf
>> libncursesada9-dev/stable 6.2.20200212-4 armhf
>> libncursesw5-dev/stable 6.2+20201114-2 armhf
>> libnotcurses-dev/stable 2.2.3+dfsg.1-1 armhf
>> librust-ncurses-dev/stable 5.99.0-3 armhf
>
> My outsider guess is that "dev" means here something different than it does for openSUSE, when I see the wording "dev/stable". They are probably experimental packages.
>
> You need help from somebody familiar with that distribution who knows what packages contain the header files, how to find them.

For a hobbyist building packages from source, "-dev"
means that is where you get your /usr/include files
for compilation.

The -dev designation means it is essential to builds.

(Not essential to application consumption by end users.
End users who do not build packages as a hobby, don't need them.)

The "pattern" for ncurses is supposed to tell you something.
Packages are likely still being build with ncurses, by
the package maintainers, so *something* in that list
is the appropriate -dev to use.

OK, so I look in Synaptic package manager:

libncurses5-dev transitional package for libncurses-dev

This is a transitional package that can safely be removed

When you select it, it installs libncurses-dev instead.

libncurses-dev

This package contains the header files, static libraries
and symbolic links that developers using ncurses will need.

There must be some logic, to the sparse matrix of packages,
but I could not justify why "this is included, that is not".
Sometimes a pattern is discern-able in the package manager,
and reading between the lines, there is a message there.

libncurses-dev

/usr/include/curses.h

/usr/lib/x86_64-linux-gnu/libncurses.so

/usr/bin/ncurses5-config \___ Another pattern. A tool for build
/usr/bin/ncurses6-config / schemes, to check for details on version

So there's a story hiding there, but I don't know what the story is.
Obviously, someone cares a great deal about some "version", to be
making such a mess :-) It's almost Pythonesque in magnitude.

Paul

Chris,

>> Question : Do you know of a technical reason why I should load the full
>> exported, rather big keymap instead of just the "wordnav.keys" one
>> (containg just the changes) ?
>
> You're loading the default full keymap and then the wordnav.keys
> OR
> You're loading the exported full keymap - which overwrites the default.
>
> Either way, whatever seems best to you.

Update : I tried both of the above, but it seems to interfere with the
'nano" editor :-(

I would like to see if, in the webpage you linked to, replacing the "Meta_b"
with the vt100 "\e[1;5C" sequence, but have no idea how to descibe the
latter. Googeling for it doesn't turn up anything. :-|

Any idea ?

Regards,
Rudy Wieser

Carlos,

> You need help from somebody familiar with that distribution who knows what
> packages contain the header files, how to find them.

Indeed.

Thanks for the reply though.

Regards,
Rudy Wieser

Paul,

> For a hobbyist building packages from source, "-dev"
> means that is where you get your /usr/include files
> for compilation.

In other words, as the above-mentioned hobbyist I should install that
"libncurses-dev" package ?

> libncurses-dev
>
> This package contains the header files, static libraries
> and symbolic links that developers using ncurses will need.

That sure does look like it. :-) May I ask where you got that "This
package contains ..." description from ?

The reason I'm asking is that I would not want to (effectivly) trash my
current installation by just dropping random packages into it.

Regards,
Rudy Wieser

On 30/04/2023 11:48, R.Wieser wrote:
> Chris,
>
>>> Question : Do you know of a technical reason why I should load the full
>>> exported, rather big keymap instead of just the "wordnav.keys" one
>>> (containg just the changes) ?
>>
>> You're loading the default full keymap and then the wordnav.keys
>> OR
>> You're loading the exported full keymap - which overwrites the default.
>>
>> Either way, whatever seems best to you.
>
> Update : I tried both of the above, but it seems to interfere with the
> 'nano" editor :-(
>
> I would like to see if, in the webpage you linked to, replacing the "Meta_b"
> with the vt100 "\e[1;5C" sequence, but have no idea how to descibe the
> latter. Googeling for it doesn't turn up anything. :-|
>
> Any idea ?
>
> Regards,
> Rudy Wieser
>
>

You started by asking how to change Ctrl-RightArrow to Word_Right *at
the terminal prompt*, similarly Ctrl-LeftArrow. The 2 keymap changes did
that by remapping to the emacs sequence word-right (left). The emacs
sequence is built in to *readline*. (You can change readline to use vi
key sequences if you want to.) So now C+-> is translated to Meta-f.

Now you've changed them (C+-> and C+<-) you can't really expect nano to
know that they've changed. Nano doesn't understand emacs sequences.

*You've changed the parameters of the question.*

Nano presumably expects C+-> to emit the *original* sequence.

NOTE: I don't completely understand the following and can't at present
test it. It needs a *proper* teminal /dev/tty? AFAICS.

You might be able to change the remapping to only affect readline by
using .inputrc (or /etc/inputrc) - i.e. remove the extra keymap file and
putting:

"\e[1~": beginning-of-line
"\e[4~": end-of-line
"\e[5~": beginning-of-history
"\e[6~": end-of-history
"\e[3~": delete-char
"\e[2~": quoted-insert
"\e[5C": forward-word
"\e[5D": backward-word
"\e\e[C": forward-word
"\e\e[D": backward-word
set completion-ignore-case On

into a file .inputrc in your home directory (or /etc/inputrc)

See here:
https://softpanorama.org/Scripting/Shellorama/Dot_files/inputrc.shtml
for example

Just as an aside, why can I find these things on the interwebs and you
can't? <g>

--
Chris Elvidge
England

On 30/04/2023 13:22, Chris Elvidge wrote:
> On 30/04/2023 11:48, R.Wieser wrote:
>> Chris,
>>
>>>> Question : Do you know of a technical reason why I should load the full
>>>> exported, rather big keymap instead of just the "wordnav.keys" one
>>>> (containg just the changes) ?
>>>
>>> You're loading the default full keymap and then the wordnav.keys
>>> OR
>>> You're loading the exported full keymap - which overwrites the default.
>>>
>>> Either way, whatever seems best to you.
>>
>> Update : I tried both of the above, but it seems to interfere with the
>> 'nano" editor :-(
>>
>> I would like to see if, in the webpage you linked to, replacing the
>> "Meta_b"
>> with the vt100 "\e[1;5C" sequence, but have no idea how to descibe the
>> latter. Googeling for it doesn't turn up anything. :-|
>>
>> Any idea ?
>>
>> Regards,
>> Rudy Wieser
>>
>>
>
> You started by asking how to change Ctrl-RightArrow to Word_Right *at
> the terminal prompt*, similarly Ctrl-LeftArrow. The 2 keymap changes did
> that by remapping to the emacs sequence word-right (left). The emacs
> sequence is built in to *readline*. (You can change readline to use vi
> key sequences if you want to.) So now C+-> is translated to Meta-f.
>
> Now you've changed them (C+-> and C+<-) you can't really expect nano to
> know that they've changed. Nano doesn't understand emacs sequences.
>
> *You've changed the parameters of the question.*
>
> Nano presumably expects C+-> to emit the *original* sequence.
>
> NOTE: I don't completely understand the following and can't at present
> test it. It needs a *proper* teminal /dev/tty? AFAICS.
>
> You might be able to change the remapping to only affect readline by
> using .inputrc (or /etc/inputrc) - i.e. remove the extra keymap file and
> putting:
>
> "\e[1~": beginning-of-line
> "\e[4~": end-of-line
> "\e[5~": beginning-of-history
> "\e[6~": end-of-history
> "\e[3~": delete-char
> "\e[2~": quoted-insert
> "\e[5C": forward-word
> "\e[5D": backward-word
> "\e\e[C": forward-word
> "\e\e[D": backward-word
> set completion-ignore-case On
>
> into a file .inputrc in your home directory (or /etc/inputrc)
>
> See here:
> https://softpanorama.org/Scripting/Shellorama/Dot_files/inputrc.shtml
> for example
>
> Just as an aside, why can I find these things on the interwebs and you
> can't? <g>
>
>

On mine, /etc/inputrc exists and says much the same as the above
..inputrc, so .inputrc could be deleted.
But, I cannot seem to make this (forward/backward word) work on console.
Meta-b and Meta-f both work properly on my console - that's what the
keymap changed.
Ctrl-RArrow and Ctrl-LeftArrow work in xfce4-terminal

--
Chris Elvidge

--
Chris Elvidge
England

Chris,

>> Update : I tried both of the above, but it seems to interfere with the
>> 'nano" editor :-(
....
> Now you've changed them (C+-> and C+<-) you can't really expect nano to
> know that they've changed. Nano doesn't understand emacs sequences.

Yes, thats also what I thought was/is happening.

> *You've changed the parameters of the question.*

From my POV I've posted a follow-up question and included the reason for
doing so : the positive change for/on the commandline has a negative effect
in the nano editor. And that means I need to find another, better solution.

My suggestion ('cause thats all it is) that I could perhaps use the official
vt100 sequence for a "word back" is something I would like to try.

> Nano presumably expects C+-> to emit the *original* sequence.

Currently both the ctrl-arrow-left and normal arrow-left emit the same
sequence, that of the normal arrow-left. and that is where my problem
started.

Personally I'm a bit surprised that nano is even able to discern between the
two keystrokes. I would not wonder if some "magic" is involved there. :-)

> You might be able to change the remapping to only affect readline by using
> .inputrc (or /etc/inputrc) - i.e. remove the extra keymap file and
> putting:
....
> "\e[5D": backward-word

:-) I do not even need to do that, as that translation (and two other
sequences doing the same) is already in the /etc/inputrc file I have. The
only thing I have to do is to make the ctrl-arrow-left generate such a
sequence.

Mind you, currently it and the normal arrow-left generate the exact same
one.

> See here:
> https://softpanorama.org/Scripting/Shellorama/Dot_files/inputrc.shtml
> for example
>
> Just as an aside, why can I find these things on the interwebs and you
> can't? <g>

You wish ! :-p

While doing my own homework I found that file too, but could not do anything
with it as long as the ctrl-arrow-left and arrow-left keypresses are
generating the same sequence. And as I could not find anything else* I
posted my question here.

* maybe I did, but didn't recognise it as such. :-\

Regards,
Rudy Wieser

On 4/30/2023 7:03 AM, R.Wieser wrote:
> Paul,
>
>> For a hobbyist building packages from source, "-dev"
>> means that is where you get your /usr/include files
>> for compilation.
>
> In other words, as the above-mentioned hobbyist I should install that
> "libncurses-dev" package ?
>
>> libncurses-dev
>>
>> This package contains the header files, static libraries
>> and symbolic links that developers using ncurses will need.
>
> That sure does look like it. :-) May I ask where you got that "This
> package contains ..." description from ?
>
> The reason I'm asking is that I would not want to (effectivly) trash my
> current installation by just dropping random packages into it.
>
> Regards,
> Rudy Wieser
>

The description comes from the Synaptic package manager window.

I installed the package in my OS. Then did "Properties" to
get a filelist of items added, in the Synaptic window.

The description is presumably a file in the .deb .

You can take a .deb apart if you want, and look at the
component parts. You don't have to install a .deb to look
at it -- .deb can also be downloaded and dissected if you prefer.

*******

The general principle, is you can turn all the packages on, without
a problem. That might even be a test case in a sense. Some packages
come with reconfigure scripts, which may ask you questions, or do
some things conditionally when setting up a package.

Adding PPA to the Package Manager, now that's on the edge (item is outside
the management of your tree), but you assume the author of your PPA
("Google Earth"), did not go out of their way to create name collisions
or overrule local policy in some way.

But the idea is, the Package Manager is not intended to be
antagonistic. It's designed to solve problems, not create them.

And Package Managers vary quite a bit, in how "friendly" they are.
In rolling release distros, where Package Management is a nightmare,
you soon learn how to use "biases", so that when the PM is faced
with impossible choices ("can't go forward, can't go back"),
application of a bias applies some arm twisting, to resolve
the issue. The Package Manager in such a case, is using human
intervention, for the desired result. Not everyone enjoys that :-)

So when you have an easy-going distro/PM, with fixed releases, and what
amount to "synchronization points", be appreciative of the misery
you are being saved. If one release used PulseAudio as a subsystem,
and the next one uses Pipewire and PulseAudio was deprecated,
then the Package Manager could enforce that, so again, any remaining
buttons in the Package Manager could be clicked, and... it all works.

Now you know why Firefox has support for three graphics options
in Linux, instead of just one. The three cases, cover situations
where the Package Management team could "kick Firefox to the curb"
if the transition support was not there. The person implementing that
at Mozilla was not happy, but.. it is what it is. I even tested,
to make sure all those options worked :-) If you use the gear wheel
during login, sometimes you can see those graphical options.

https://linuxconfig.org/wp-content/uploads/2020/04/01-how-to-enable-disable-wayland-on-ubuntu-20-04-desktop.png

If you do that today, there could still be rough edges. This stuff
has been ten years in the making. There are bound to be some
very nice things, that will be lost because of this.

Paul

Paul,

>> That sure does look like it. :-) May I ask where you got that "This
>> package contains ..." description from ?
....
> The description comes from the Synaptic package manager window.

Ah. I have to do with "apt search" I'm afraid (am trying to keep my
"bullseye lite" /lite/ )

> You can take a .deb apart if you want, and look at the
> component parts.

:-) Thats something I was already considering, though for another reason.

Regards,
Rudy Wieser

Chris,

> I would like to see if, in the webpage you linked to, replacing the
> "Meta_b" with the vt100 "\e[1;5C" sequence, but have no idea how to
> descibe the latter. Googeling for it doesn't turn up anything. :-|

Although my question was aimed at which "magic phrase" I would have needed
instead of "Meta_b" to get the above vt100 sequence (I've not been able to
find a list of possible "magic phrases") I realized that I have already come
across something which could do it. After a bit of back-reading and testing
and the below is now the result.

- the Keymap file for loadkeys

control keycode 105 = F70
string F70 = "\033[1;5D"
control keycode 106 = F71
string F71 = "\033[1;5C"

It seems to work for both the commandline as well as the nano editor.

Regards,
Rudy Wieser

On 01/05/2023 07:02, R.Wieser wrote:
> Chris,
>
>> I would like to see if, in the webpage you linked to, replacing the
>> "Meta_b" with the vt100 "\e[1;5C" sequence, but have no idea how to
>> descibe the latter. Googeling for it doesn't turn up anything. :-|
>
> Although my question was aimed at which "magic phrase" I would have needed
> instead of "Meta_b" to get the above vt100 sequence (I've not been able to
> find a list of possible "magic phrases") I realized that I have already come
> across something which could do it. After a bit of back-reading and testing
> and the below is now the result.
>
> - the Keymap file for loadkeys
>
> control keycode 105 = F70
> string F70 = "\033[1;5D"
> control keycode 106 = F71
> string F71 = "\033[1;5C"
>
> It seems to work for both the commandline as well as the nano editor.
>
>
> Regards,
> Rudy Wieser
>
>

OK, good.

--
Chris Elvidge
England