From: Mark Jamsek <mark@jamsek.com>
Subject: Re: [rfc] compound keymaps with prefixed count modifier like vi(1)
To: Christian Weisgerber <naddy@mips.inka.de>
Cc: gameoftrees@openbsd.org
Date: Thu, 23 Jun 2022 01:44:03 +1000

On 22-06-22 05:33pm, Christian Weisgerber wrote:
> Omar Polo:
> 
> > abort keys are IMHO one of the weakest point of tog :)
> 
> Indeed.
> 
> > If I can suggest something here, I'd go with with C-g.  I'm biased
> > because Emacs uses it to interrupt various kinds of things, but less(1)
> > uses it too.  ESC maybe would be familiar for vi users?
> 
> I'm also partial to ^G.  (Mutt is another application that uses it.)
> 
> ESC is a poor choice.  Since the character introduces all sorts of
> special sequences, like those for cursor keys, a single ESC is
> difficult to separate from ESC as part of a sequence.  I think
> curses mostly does this by way of a timeout, i.e., if no other
> character follows for 500 ms, a single ESC is recognized as such.
> IIRC, you can also type ESC-ESC to speed this up.  It's cumbersome
> and error-prone, and then you have to lecture users how character-based
> terminals work, and...
> 
> Basically, try to avoid using ESC for anything in a curses-based
> user interface.

Thanks for the tip!

That explains what I was seeing in my testing earlier. I have another
diff in this thread where we use ESC and C-g to abort operations, and it
felt like ESC was taking longer than C-g.

I think it would be a good idea to standardise on C-g and perhaps
backspace as the abort sequence in tog.

-- 
Mark Jamsek <fnc.bsdbox.org>
GPG: F2FF 13DE 6A06 C471 CA80  E6E2 2930 DC66 86EE CF68