Could anyone tackle Emacs bug#23794 in sort-lines before Emacs 26 is released?

Could anyone tackle Emacs bug#23794 in sort-lines before Emacs 26 is released?

[Robert Weiner]

[-- Attachment #1: Type: text/plain, Size: 1021 bytes --]

This is the bug that sort-lines since Emacs 25 has not handled collapsed
(invisible) lines properly (keeping them with their parents) when sorting.
Prior to Emacs 25, sort-lines handled outline sorting properly.

It should be a fairly straightforward thing (relative to the complexity of
other Emacs issues) since it is well understood and a potential patch has
been offered.  Eli suggested that any change be limited to outlining modes
to limit its effects.

​I cannot work on it myself right now and that is why I am asking someone
else to pick it up.​


I think this is a bug that should be eliminated before Emacs 26 goes live
​ so it does not impact people with large outlines who might sort them and
get the wrong results but not notice until their data has been corrupted
for a while and they can't get back to the original proper hierarchical
associations.​

At the very least the present behavior makes no sense for outline modes
that use invisibility for collapsing.

​Bob
​

[-- Attachment #2: Type: text/html, Size: 1994 bytes --]

Re: Could anyone tackle Emacs bug#23794 in sort-lines before Emacs 26 is released?

[John Wiegley]

>>>>> "RW" == Robert Weiner <rsw@gnu.org> writes:

RW> I think this is a bug that should be eliminated before Emacs 26 goes live
RW> so it does not impact people with large outlines who might sort them and
RW> get the wrong results but not notice until their data has been corrupted
RW> for a while and they can't get back to the original proper hierarchical
RW> associations.​

This does seem like something worth fixing in 26. Do you have a sense of how
large a change this might be? Is anyone else willing to take up the task?

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: Could anyone tackle Emacs bug#23794 in sort-lines before Emacs 26 is released?

[Eli Zaretskii]

> From: John Wiegley <jwiegley@gmail.com>
> Date: Thu, 21 Sep 2017 22:59:54 -0700
> Cc: rswgnu@gmail.com, emacs-devel <emacs-devel@gnu.org>
> 
> >>>>> "RW" == Robert Weiner <rsw@gnu.org> writes:
> 
> RW> I think this is a bug that should be eliminated before Emacs 26 goes live
> RW> so it does not impact people with large outlines who might sort them and
> RW> get the wrong results but not notice until their data has been corrupted
> RW> for a while and they can't get back to the original proper hierarchical
> RW> associations.​
> 
> This does seem like something worth fixing in 26. Do you have a sense of how
> large a change this might be?

You can see it here:

  https://debbugs.gnu.org/cgi/bugreport.cgi?bug=23794#20

It lacks documentation bits.  And I'd be glad to see someone else
commenting on the problem and the solution.

Btw: the problem was not introduced in Emacs 25, as Robert says, it
was introduced much earlier, in Emacs 22.