bug#60467: 30.0.50; primitive-undo: Changes to be undone by function different from announced

[Gregory Heytings]

> > > For the loop currently in `subr.el` I think the (cdr ptr) is OK 
> > > because we use the `setcdr` to truncate the list, so that last cons 
> > > (whose `cdr` may be nil) will be in the list included in the (apply 
> > > ....) entry.
> >
> > Indeed, now I see it, and I think (without being 100% sure) that it's 
> > OK. It's yet another obscure point of the original code, though.
>
> I hope all these revelations will find their way into comments to this 
> tricky code.

Well, I hope this tricky code will be replaced by the non-tricky code I 
proposed, in which none of these obscurities exist... ;-)