On Thu, Feb 26, 2026 at 09:02:52AM +0100, Dr. Werner Fink wrote:
> On 2026/02/25 16:06:20 -0500, Thomas Dickey wrote:
> > On Wed, Feb 25, 2026 at 01:58:28PM +0100, Dr. Werner Fink wrote:
> > > Hi,
> > > 
> > > I've a bug report about dialog not supporting the Alt key. Means
> > > e.g. that
> > > 
> > >    dialog --yesno Hallo 6 30
> > > 
> > > return for both Alt-y and Alt-n or Alt-Y and Alt-N the value 255.
> > > This is nowadays somehow uncommon[1].  It would very helpful 
> > > to be able to navigate not only with cursor or tab keys but
> > > also with the ALt and the coloured characters of the buttons.
> > 
> > dialog uses the terminfo description.  It happens to know about xterm's
> > alternate screen:
> > 
> >        --keep-tite
> > 
> > Generally (no counter examples come to mind)...
> > 
> >     + what some people expect with alt+whatever
> 
> Most graphical apps do use Alt-<Char> to press a Button with <Char>
> in its label. Also the old YaST does this in its graphical UI as well
> as in its ncurses UI.
> 
> And *some* people are the most normal user out there ;)

A "normal user" is someone who uses a browser to interact with the computer.

But that's not what dialog is doing.

The normal user of dialog isn't using zenity (except as a workaround).
 
> >     + is a corruption of meta mode (apparently due to bash),
> 
> Hmm .. I'm using tcsh as well as the bash, sometime ksh ... in all
> cases the XTerm shows for Alt-<Char> the ESC-<Char> sequence ...
> even in clisp I see ESC-<Char>

sure - it was a Linux console "feature" in the mid-1990s.
 
> With enable-meta-key set to On remains in ESC-<Char> ...only
> disabling "Alt sends Escape" in XTerm change that as expected.
> 
> Beside this the problem occurs on the Linux console
> 
> > 
> >     + that would magically send an escape character before _some_ keys
> >       (more often before keys that aren't special such as cursor
> >       and function keys -- but that assumption isn't reliable)
> > 
> >     + which isn't related to the terminfo description
> >       (recalling that bash's developer(s) prefer hardcoding to extending
> >       termcap or terminfo).
> > 
> > So... it's not in the terminfo description, and any attempt to make it
> > a "standard" feature would run into the exceptions to the normal/special
> > keys assumption.
> > 
> >     + the place to start with a new feature would be with the developers
> >       of terminals and/or shells, etc., who would define it (and agree
> >       on what's common behavior).
> >  
> > > 
> > > Werner
> > > 
> > > [1] https://bugzilla.opensuse.org/show_bug.cgi?id=1257107

I don't see any proposed contributions in that bug report - just a lot
of "I want".

See this for the context in which those are received:

https://invisible-island.net/personal/self-service.html

-- 
Thomas E. Dickey <[email protected]>
https://invisible-island.net

Attachment: signature.asc
Description: PGP signature

Reply via email to