Re: [Patch]: Allow overlay arrows to be inserted before the text at column zero rather than splatting it.

[Alan Mackenzie <acm@muc.de>]

Hello, Eli.

On Mon, Aug 19, 2019 at 05:29:42 +0300, Eli Zaretskii wrote:
> > Date: Sun, 18 Aug 2019 19:23:10 +0000
> > Cc: emacs-devel@gnu.org
> > From: Alan Mackenzie <acm@muc.de>

> > > Sorry, I'm confused now: how can you insert the arrow without either
> > > moving the rest of the line to the right or overwriting its first few
> > > glyphs?  Where does the space for the arrow come from?

> > Sorry for being a bit inexact.  In that bit I meant inserting "=>" by
> > overwriting the two characters in columns 0 and 1.  We need to be able
> > to do this.

> But AFAIR that's exactly what the current implementation of
> overlay-arrow on TTY frames does.

It is.  I was talking here about the putative superseding of the overlay
arrow with (?higher level?) code using the display property in place of
the overlay arrow's current, ugly, mechanism.

It seems I switched over the subject I was talking about without making
this clear.  Sorry.  The subject I started with at the beginning of the
thread was extending the overlay arrow mechanism with my patch to
xdisp.c.  I can understand people not liking this patch.

> So how exactly is your implementation different in this regard, and
> why did you need any new code for it?

The patch I proposed would have allowed the overlay arrow's mechanism to
be used to _insert_ "=>" before column zero on a line (rather than
overwriting the first two characters).

The general feeling in this thread is that I should rethink this patch,
instead using before-string overlay properties.  This would enable me to
close bug #36832 (The one about not wanting to scroll the compilation
mode buffer on typing CR) satisfactorally.

I intend to do this in the coming days.

> Still confused.

I'll try to be a bit more considerate in future threads, when my
thinking starts getting unstable, and there's a danger of me zipping
between confusingly related, but different, topics.

I hope this thread has become at least moderately clear, now.

-- 
Alan Mackenzie (Nuremberg, Germany).