On Wed, Jul 17, 2024 at 03:54:12PM +0200, Walter Alejandro Iglesias wrote:
> Hi Rene,
>
> First of all, thank you for being considerate of me. :-)
>
> On Wed, Jul 17, 2024 at 02:22:33PM +0200, Rene Kita wrote:
> > On Wed, Jul 17, 2024 at 12:59:48PM +0200, Walter Alejandro Iglesias wrote:
> > > Excuse my insistence, but I use vi-like bindings even in my shell, after
> > > using mutt for decades I still insult RMS every time I mistakenly press
> > > Esc to abort a command. :-)
> > >
> > > Could you give my patches *one* chance or at least let me know why they
> > > are ignored, please?
> > >
> > >    https://en.roquesor.com/Downloads/muttesckey.tar.gz
> >
> > While I can't say why your patches are ignored, I can say that they will
> > not be accepted until Kevin changes his mind or we find a solution on
> > the way towards a v2.3, see this[0] mail from two years ago.
> >
> > I had a very quick look at your proposed patches and my first impression
> > was, that adding another special case for one key is not the correct
> > approach for a patch going mainline. It's reasonable as a patch, but as
> > a general solution I think is preferable to have an option to configure
> > an abort key. Some people might prefer ^C to abort or some other key.
>
> I think you're confusing (shell) cancel with abort.  Ctrl-C is also
> there.

No, I'm not confusing it. Ctrl-C came just to my mind - maybe because
I'm also a w3m user. I regularly use Ctrl-C Ctrl-G to cancel something
in mutt. ;-)

> Besides, giving the user an option to choose any key binding to
> abort is difficult to accomplish for the way this special binding was
> hardcoded in mutt.  It's not *my approach*, is the way mutt has been
> designed and coded.  I didn't investigate it enough but I remember
> reading (please correct me if I'm wrong) that it's ncurses what narrows
> the choice with this particular action.  Notice that it's not only mutt,
> all ncurses applications are flexible in letting you choose key bindings
> for everything except for abort.

OK, I'm not really familiar with this topic. I just wanted to start a
discussion and that was something that could be a reason to not pick a
patch. I'm aware that such changes in old code are often more
complicated than expected.

> Personally I modify or customize my system only when it's absolutely
> necessary, I build on what's already done, I don't try to reinvent the
> wheel.  Emacs and vi are the traditional editing modes, I chose vi
> because as a BSD user I found it less problematic (the meta key doesn't
> always work, see in muttrc(5) the "meta_key" option).  So far mutt
> forces you to use Ctrl-G (emacs-like binding), with my patches you also
> have the vi-like binding version.

As a fellow vi user I appreciate this and am going to test your patches,
thanks.

> If all ncurses-based applications let you do this (w3m comes also to
> mind) I would be more than satisfied.

As a contributor to w3m I'd be interested to here what you are missing.
Feel free to tell me in private.

> Think about how much work it would relieve the maintainer (one example
> among many) to let SMTP, IMAP and POP client functionality to software
> specifically designed, created and maintained for that purpose.  This is
> not *my approach* either, this is Unix philosophy.

That ship might have sailed - but I totally agree. It might be worth
thinking about a fork without all this, but if you remove SMTP, IMAP and
POP, is it still a mutt?

Reply via email to