Re: gnus-change-server problems?

[Steven L Baur <steve@xemacs.org>]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Chris Halverson <cdh@CompleteIS.com> writes:
>> > (sort '(nil) '<)
>> > 
>> > Put point after the closing paren.  Press `C-x C-e'.  Does your XEmacs
>> > say "beep" or "(nil)"?
>> 
>> (nil)

> But then I don't understand the backtrace you posted:

> Signaling: (wrong-type-argument number-or-marker-p nil)
>   <(nil nil)
>   sort((nil) <)
>   gnus-move-group-to-server(("alt.collecting.pens-pencils" 3 ((1 . 9531)) ((save 9486) (expire 8963 ... 9126 9130 9133 9142 9342 ... ...) (reply 9300 9486 9513))) (nntp "news.mr.net") (nntp "news.uslink.net"))

> Am I misreading this backtrace?  Doesn't it say that `(sort '(nil) '<)' 
> bugs out?

That's how I read it too.  On the other hand, the sort subr hasn't
been changed in a long time (the code dates from the Lucid days), and
it should be impossible to hit any code where a comparison is done
when it is handed a nil list.  (< nil nil) definitely produces the
same wrong-type-argument error message shown in the backtrace.

Chris, can you possibly manage to get a C backtrace of this error?

Note that the code that catches degenerate lists in XEmacs' list_sort
is identical to the in-line code in Emacs 19.34+ sort.

DEFUN ("sort", Fsort, 2, 2, 0, /*
Sort LIST, stably, comparing elements using PREDICATE.
Returns the sorted list.  LIST is modified by side effects.
PREDICATE is called with two elements of LIST, and should return T
if the first element is \"less\" than the second.
*/
       (list, pred))
{
  return list_sort (list, pred, merge_pred_function);
}

Lisp_Object
list_sort (Lisp_Object list,
           Lisp_Object lisp_arg, 
           int (*pred_fn) (Lisp_Object, Lisp_Object,
                           Lisp_Object lisp_arg))
{
  Lisp_Object front, back;
  Lisp_Object len, tem;
  struct gcpro gcpro1, gcpro2, gcpro3;
  int length;

  front = list;
  len = Flength (list);
  length = XINT (len);
  if (length < 2)
    return list;            <====== A nil list should be caught here.
  ...
-- 
steve@calag.com baur
Unsolicited commercial e-mail will be billed at $250/message.