GNU bug report logs - #78965
31.0.50; exchange-point-and-mark can create duplicate adjacent entries in mark-ring

Previous Next

To reply to this bug, email your comments to 78965 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Jake <jforst.mailman <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: 31.0.50; exchange-point-and-mark can create duplicate adjacent
 entries in mark-ring
Date: Sun, 6 Jul 2025 03:09:26 +0000

Hello

If you have >= 1 mark in the local mark ring, use `set-mark-command`
with prefix argument ("C-u C-SPC") and then `exchange-point-and-mark`
("C-x C-x"), mark is equal to the last element in the local mark ring.
If you then "C-u C-SPC", the last 2 elements in the local mark ring
contain the same marker position.

Is that useful for anything?  It seems suboptimal when cycling through
the mark ring.  I think it would be similar to if `set-mark-command`
allowed the same marker position to be repeatedly pushed onto the mark
ring, which it doesn't.

Any thoughts?

The below patch seems to address the issue.

Thanks
Jake

diff --git a/lisp/simple.el b/lisp/simple.el
index fa173b26289..1da9c816ef3 100644
--- a/lisp/simple.el
+++ b/lisp/simple.el
@@ -7531,6 +7531,8 @@ temporarily."
     (if (null omark)
         (user-error "No mark set in this buffer"))
     (set-mark (point))
+    (if (equal (mark) (marker-position (car (last mark-ring))))
+        (setq mark-ring (butlast mark-ring)))
     (goto-char omark)
     (cond (temp-highlight)
           ((xor arg (if exchange-point-and-mark-highlight-region

Message #8 received at 78965 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Jake <jforst.mailman <at> gmail.com>
Cc: 78965 <at> debbugs.gnu.org
Subject: Re: bug#78965: 31.0.50;
 exchange-point-and-mark can create duplicate adjacent entries in
 mark-ring
Date: Sun, 06 Jul 2025 08:09:53 +0300

> From: Jake <jforst.mailman <at> gmail.com>
> Date: Sun, 6 Jul 2025 03:09:26 +0000
> 
> If you have >= 1 mark in the local mark ring, use `set-mark-command`
> with prefix argument ("C-u C-SPC") and then `exchange-point-and-mark`
> ("C-x C-x"), mark is equal to the last element in the local mark ring.
> If you then "C-u C-SPC", the last 2 elements in the local mark ring
> contain the same marker position.
> 
> Is that useful for anything?  It seems suboptimal when cycling through
> the mark ring.

How is it suboptimal?  Can you show a recipe, from "emacs -Q", which
illustrates the suboptimal nature of this?

> I think it would be similar to if `set-mark-command`
> allowed the same marker position to be repeatedly pushed onto the mark
> ring, which it doesn't.
> 
> Any thoughts?

Since this is very old behavior, I'd prefer not to touch it, unless it
shows a clear bug in  some case.  Who knows what optional features out
there could rely on this?

Thanks.

Message #11 received at 78965 <at> debbugs.gnu.org (full text, mbox):

From: Jake <jforst.mailman <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 78965 <at> debbugs.gnu.org
Subject: Re: bug#78965: 31.0.50; exchange-point-and-mark can create duplicate
 adjacent entries in mark-ring
Date: Sun, 6 Jul 2025 06:51:19 +0000

Hi Eli

> How is it suboptimal?

Having duplicate *adjacent* entries in the mark ring could be
suboptimal for a couple of reasons:
1. When navigating back through the mark ring, you have to do an extra
C-u C-SPC to get to the next unique position.
2. The mark ring is supposed to store valuable/significant/important
positions and it has finite size (mark-ring-max is 16 by default), so
having an adjacent duplicate could take up valuable space in the mark
ring.

From emacs -Q,
- C-SPC C-SPC
- C-p
- C-SPC C-SPC
- C-p
- C-SPC C-SPC (mark ring now has 2 unique elements)
- C-u C-SPC
- C-x C-x
- C-u C-SPC (mark ring now has 2 elements and they are equal)

Also, I think this might be a better and more expressive patch than my
previous one:

diff --git a/lisp/simple.el b/lisp/simple.el
index fa173b26289..486901cea9d 100644
--- a/lisp/simple.el
+++ b/lisp/simple.el
@@ -7490,7 +7490,9 @@ In Transient Mark mode, activate mark if
optional third arg ACTIVATE non-nil."
   "Pop off mark ring into the buffer's actual mark.
 Does not set point.  Does nothing if mark ring is empty."
   (when mark-ring
-    (setq mark-ring (nconc mark-ring (list (copy-marker (mark-marker)))))
+    (unless (equal (marker-position (mark-marker))
+                   (marker-position (car (last mark-ring))))
+      (setq mark-ring (nconc mark-ring (list (copy-marker (mark-marker))))))
     (set-marker (mark-marker) (car mark-ring))
     (set-marker (car mark-ring) nil)
     (pop mark-ring))

On Sun, Jul 6, 2025 at 5:10 AM Eli Zaretskii <eliz <at> gnu.org> wrote:
>
> > From: Jake <jforst.mailman <at> gmail.com>
> > Date: Sun, 6 Jul 2025 03:09:26 +0000
> >
> > If you have >= 1 mark in the local mark ring, use `set-mark-command`
> > with prefix argument ("C-u C-SPC") and then `exchange-point-and-mark`
> > ("C-x C-x"), mark is equal to the last element in the local mark ring.
> > If you then "C-u C-SPC", the last 2 elements in the local mark ring
> > contain the same marker position.
> >
> > Is that useful for anything?  It seems suboptimal when cycling through
> > the mark ring.
>
> How is it suboptimal?  Can you show a recipe, from "emacs -Q", which
> illustrates the suboptimal nature of this?
>
> > I think it would be similar to if `set-mark-command`
> > allowed the same marker position to be repeatedly pushed onto the mark
> > ring, which it doesn't.
> >
> > Any thoughts?
>
> Since this is very old behavior, I'd prefer not to touch it, unless it
> shows a clear bug in  some case.  Who knows what optional features out
> there could rely on this?
>
> Thanks.

Message #14 received at 78965 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Jake <jforst.mailman <at> gmail.com>
Cc: 78965 <at> debbugs.gnu.org
Subject: Re: bug#78965: 31.0.50; exchange-point-and-mark can create duplicate
 adjacent entries in mark-ring
Date: Sun, 06 Jul 2025 10:04:28 +0300

> From: Jake <jforst.mailman <at> gmail.com>
> Date: Sun, 6 Jul 2025 06:51:19 +0000
> Cc: 78965 <at> debbugs.gnu.org
> 
> > How is it suboptimal?
> 
> Having duplicate *adjacent* entries in the mark ring could be
> suboptimal for a couple of reasons:
> 1. When navigating back through the mark ring, you have to do an extra
> C-u C-SPC to get to the next unique position.

IMO, it's a minor annoyance in a relatively rare use case.

> 2. The mark ring is supposed to store valuable/significant/important
> positions and it has finite size (mark-ring-max is 16 by default), so
> having an adjacent duplicate could take up valuable space in the mark
> ring.

You can enlarge the ring size if you have problems with the default
size.

And I don't think I agree with the "valuable/significant/important"
part.

> > Since this is very old behavior, I'd prefer not to touch it, unless it
> > shows a clear bug in  some case.  Who knows what optional features out
> > there could rely on this?

This is still a prevailing consideration, from where I stand.

Message #17 received at 78965 <at> debbugs.gnu.org (full text, mbox):

From: Jake <jforst.mailman <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 78965 <at> debbugs.gnu.org
Subject: Re: bug#78965: 31.0.50; exchange-point-and-mark can create duplicate
 adjacent entries in mark-ring
Date: Sun, 6 Jul 2025 07:18:01 +0000

> You can enlarge the ring size if you have problems with the default
size.

Someone who likes to navigate with the mark-ring would be more likely
to customise mark-ring-max to be smaller, so they get back to where
they started quicker.  In that case, space in the mark ring becomes
more valuable, and this issue becomes more of a problem.

Having said that, I agree that it's a minor issue, so I understand if
you don't want to change long-standing behaviour or introduce an
option for it.

Thanks again
Jake

On Sun, Jul 6, 2025 at 7:04 AM Eli Zaretskii <eliz <at> gnu.org> wrote:
>
> > From: Jake <jforst.mailman <at> gmail.com>
> > Date: Sun, 6 Jul 2025 06:51:19 +0000
> > Cc: 78965 <at> debbugs.gnu.org
> >
> > > How is it suboptimal?
> >
> > Having duplicate *adjacent* entries in the mark ring could be
> > suboptimal for a couple of reasons:
> > 1. When navigating back through the mark ring, you have to do an extra
> > C-u C-SPC to get to the next unique position.
>
> IMO, it's a minor annoyance in a relatively rare use case.
>
> > 2. The mark ring is supposed to store valuable/significant/important
> > positions and it has finite size (mark-ring-max is 16 by default), so
> > having an adjacent duplicate could take up valuable space in the mark
> > ring.
>
> You can enlarge the ring size if you have problems with the default
> size.
>
> And I don't think I agree with the "valuable/significant/important"
> part.
>
> > > Since this is very old behavior, I'd prefer not to touch it, unless it
> > > shows a clear bug in  some case.  Who knows what optional features out
> > > there could rely on this?
>
> This is still a prevailing consideration, from where I stand.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.