Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

[Stefan Monnier]

> +	(if (< n 0)
> +	    (move-point-visually -1)
> +	  (move-point-visually 1))

Aka
	(move-point-visually (if (< n 0) -1 1))

> +	(sit-for 0))

I think this deserves a comment, because I for one have no idea why
it's here.


        Stefan

Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: emacs-devel@gnu.org
> Date: Sat, 29 Jun 2013 11:12:34 -0400
> 
> > +	(if (< n 0)
> > +	    (move-point-visually -1)
> > +	  (move-point-visually 1))
> 
> Aka
> 	(move-point-visually (if (< n 0) -1 1))

Fixed.

> > +	(sit-for 0))
> 
> I think this deserves a comment, because I for one have no idea why
> it's here.

It's probably my misunderstanding of something.  If I remove sit-for,
invoking the command with N > 1 causes it to see the value of point
that is not updated.  That is, if point is at position X and you
invoke

  C-u 10 <right>

then inside move-point-visually I see on the 2nd or 3rd call that
point is being reset back to its value before the first call.

Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

[Eli Zaretskii]

> Date: Sat, 29 Jun 2013 18:24:29 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> CC: emacs-devel@gnu.org
> 
> > > +	(sit-for 0))
> > 
> > I think this deserves a comment, because I for one have no idea why
> > it's here.
> 
> It's probably my misunderstanding of something.  If I remove sit-for,
> invoking the command with N > 1 causes it to see the value of point
> that is not updated.  That is, if point is at position X and you
> invoke
> 
>   C-u 10 <right>
> 
> then inside move-point-visually I see on the 2nd or 3rd call that
> point is being reset back to its value before the first call.

I found the reason: it was a bug.  The code moved point by using
information in the current glyph matrix, but did not invalidate the
cursor position info in the matrix, so redisplay would not update the
cursor position; calling sit-for forced a more thorough redisplay and
"fixed" that.

I fixed the code by invalidating the cursor location, and removed the
call to sit-for.

Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

[Stefan Monnier]

> The code moved point by using
> information in the current glyph matrix,

Does that means that those functions won't work in batch mode?

> I fixed the code by invalidating the cursor location, and removed the
> call to sit-for.

Great, thanks,


        Stefan

Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

[Eli Zaretskii]

> From: Stefan Monnier <monnier@IRO.UMontreal.CA>
> Cc: emacs-devel@gnu.org
> Date: Sun, 30 Jun 2013 17:33:37 -0400
> 
> > The code moved point by using
> > information in the current glyph matrix,
> 
> Does that means that those functions won't work in batch mode?

They should still work, because there's fallback code which uses
move_it_* functions (needed when the position to the left or right is
outside of the currently displayed portion of the buffer).  The code
that uses the current glyph matrices was written because it is
expected to be used almost always in interactive invocation of the
function, since moving point almost never invalidates the current
glyph matrix, and the resulting display update shortcuts almost
everything.

Re: [Emacs-diffs] trunk r113221: Implement visual-order cursor motion.

[Stefan Monnier]

> They should still work, because there's fallback code which uses
> move_it_* functions (needed when the position to the left or right is
> outside of the currently displayed portion of the buffer).  The code
> that uses the current glyph matrices was written because it is
> expected to be used almost always in interactive invocation of the
> function, since moving point almost never invalidates the current
> glyph matrix, and the resulting display update shortcuts almost
> everything.

OK, great,


        Stefan