bug#15059: 24.3; nth with negative index values

bug#15059: 24.3; nth with negative index values

[Wilfred Hughes]

I've noticed that nth returns the first item in a list, when it's given
a negative index.

(nth -1 '(5 6 7)) ;; 5
(nth -2 '(5 6 7)) ;; 5

Judging by my reading of the docstring, I was expecting nil (though
raising an error might be a sensible behaviour instead).

In GNU Emacs 24.3.1 (x86_64-unknown-linux-gnu, GTK+ Version 3.8.2)
 of 2013-07-30 on -var-lib-archbuild-staging-x86_64-jgc
Windowing system distributor `The X.Org Foundation', version 11.0.11402000
Configured using:
 `configure '--prefix=/usr' '--sysconfdir=/etc' '--libexecdir=/usr/lib'
 '--localstatedir=/var' '--with-x-toolkit=gtk3' '--with-xft'
 'CFLAGS=-march=x86-64 -mtune=generic -O2 -pipe -fstack-protector
 --param=ssp-buffer-size=4'
 'LDFLAGS=-Wl,-O1,--sort-common,--as-needed,-z,relro'
 'CPPFLAGS=-D_FORTIFY_SOURCE=2''

Important settings:
  value of $LANG: en_GB.utf8
  value of $XMODIFIERS: @im=ibus
  locale-coding-system: utf-8-unix
  default enable-multibyte-characters: t

bug#15059: 24.3; nth with negative index values

[Andreas Schwab]

Wilfred Hughes <wilfred@editd.com> writes:

> I've noticed that nth returns the first item in a list, when it's given
> a negative index.
>
> (nth -1 '(5 6 7)) ;; 5
> (nth -2 '(5 6 7)) ;; 5
>
> Judging by my reading of the docstring, I was expecting nil

The list is longer than -1, so this case doesn't apply.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

bug#15059: 24.3; nth with negative index values

[Glenn Morris]

Wilfred Hughes wrote:

> (nth -1 '(5 6 7)) ;; 5
> (nth -2 '(5 6 7)) ;; 5
>
> Judging by my reading of the docstring, I was expecting nil (though
> raising an error might be a sensible behaviour instead).

Easy to raise an error (unless something was relying on the behaviour
for -ve arguments?):

*** src/fns.c	2013-08-01 22:24:02 +0000
--- src/fns.c	2013-08-09 18:43:02 +0000
***************
*** 1277,1283 ****
    (Lisp_Object n, Lisp_Object list)
  {
    EMACS_INT i, num;
!   CHECK_NUMBER (n);
    num = XINT (n);
    for (i = 0; i < num && !NILP (list); i++)
      {
--- 1277,1283 ----
    (Lisp_Object n, Lisp_Object list)
  {
    EMACS_INT i, num;
!   CHECK_NATNUM (n);
    num = XINT (n);
    for (i = 0; i < num && !NILP (list); i++)
      {

bug#15059: 24.3; nth with negative index values

[Stefan Monnier]

> Easy to raise an error (unless something was relying on the behaviour
> for -ve arguments?):

We could try.  But note that Emacs-19.34 already behaved like that (I
don't have any older version at hand to test).  Not sure it's important
to "fix".


        Stefan

bug#15059: 24.3; nth with negative index values

[Glenn Morris]

Stefan Monnier wrote:

>> Easy to raise an error (unless something was relying on the behaviour
>> for -ve arguments?):
>
> We could try.  But note that Emacs-19.34 already behaved like that (I
> don't have any older version at hand to test).  Not sure it's important
> to "fix".

Oh, the elisp manual actually documents the -ve argument case:

   -- Function: nth n list
  [...]
  If N is negative, `nth' returns the first element of LIST.
  [...]
  (nth -3 '(1 2 3 4))
               => 1

All hope is lost!

bug#15059: 24.3; nth with negative index values

[Wilfred Hughes]

[-- Attachment #1: Type: text/plain, Size: 160 bytes --]

Ah, great to see this behaviour is documented. Could the docstring of nth
and nthcdr be changed to also point to this part of the elisp manual,
similar to cdr?

[-- Attachment #2: Type: text/html, Size: 164 bytes --]

bug#15059: 24.3; nth with negative index values

[Stefan Monnier]

> Ah, great to see this behaviour is documented. Could the docstring of nth
> and nthcdr be changed to also point to this part of the elisp manual,
> similar to cdr?

I prefer to leave it undocumented where possible (I'd be even tempted
to remove the part of the Elisp manual that documents it).
Just don't rely on this behavior, it's a bad idea.


        Stefan

bug#15059: 24.3; nth with negative index values

[Glenn Morris]

Stefan Monnier wrote:

> (I'd be even tempted to remove the part of the Elisp manual that
> documents it).

I was thinking the same thing, so I did that.