Re: Fwd: Re: Strange problem with bindkey keys passing through to apps

[David Woodfall <dave@dawoodfall.net>]

On Friday 14 September 2018 15:32,
Daniel Shahaf <danielsh@apache.org> put forth the proposition:

> Resending again..

Thanks. I've reverted back to the opendns servers until I can resolve
this issue. It seems that any ipv6 lookups are coming back as a
non-existent domain.

> Daniel Shahaf wrote on Fri, 14 Sep 2018 14:54 +0000:
> > David Woodfall wrote on Fri, 14 Sep 2018 14:23 +0100:
> > > I did do try in screen but found that e.g. vim and weechat still see
> > > it. I think I need to look into application level keys for it, rather
> > > than a normal bind.
> >
> > In screen the binding would be Ctrl-A Alt-s, not bare Alt-s.
> >
> > Daniel

It's possible to add binds that work without needing to press ctrl-a
though. If you enter the screen prompt with 'Ctrl-a :' and then enter
the command bindkey with an option it will show a few types of keys
bindings. 

With -a or no option it will show user keys. It's possible to set
other key combos for 'command' (that is usually Ctrl-a.) e.g. I added
Alt-a too because I find that easier on my current keyboard, although
I left the default Ctrl-a in for safety:

bindkey ^[a command

The ones labelled [A] with the -d flag are application level that
also don't need ctrl-a first. But care must be taken with those.
The ones labelled :: are arrow keys and such and are read from the
terminfo or termcap.

I've tested Alt-s:

bindkey ^[s exec printf "\eP\e[15]\e\\"

and it seems to work in mutt within screen at least. I haven't tried
anything else yet.

--

If money can't buy happiness, I guess you'll just have to rent it.

                                                            .--.  oo
                                                           (____)//
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~'