From xemacs-m  Thu Sep 18 16:12:49 1997
Received: from pocari-sweat.jprc.com (POCARI-SWEAT.JPRC.COM [207.86.147.217])
	by xemacs.org (8.8.5/8.8.5) with ESMTP id QAA10864
	for <xemacs-beta@xemacs.org>; Thu, 18 Sep 1997 16:12:47 -0500 (CDT)
Received: (from karl@localhost)
	by pocari-sweat.jprc.com (8.8.5/8.8.5) id RAA08003;
	Thu, 18 Sep 1997 17:11:45 -0400
To: ding@gnus.org
Cc: xemacs-beta@xemacs.org
Subject: Re: [XEmacs] Re: Quassia Gnus v0.7 is released
References: <m2en6ngwhn.fsf@proletcult.slip.ifi.uio.no> <m2afhbkwnm.fsf@altair.xemacs.org> <m2sov2l8fq.fsf@altair.xemacs.org>
From: Karl Kleinpaste <karl@jprc.com>
Date: 18 Sep 1997 17:11:44 -0400
In-Reply-To: SL Baur's message of "18 Sep 1997 12:37:45 -0700"
Message-ID: <vxkhgbie38v.fsf@pocari-sweat.jprc.com>

SL Baur <steve@xemacs.org> writes:
> I managed to get a lisp backtrace.  XEmacs should not be dropping
> core, but it looks like something weird is going on with nndraft.

>   directory-files("/home/steve/News/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/cache/news.admin.net-abuse.announce" nil "^[0-9]+$" t)
>   # bind (dir)
>   nnmh-request-list-1("/home/steve/News/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/cache/news.admin.net-abuse.announce")
>   # bind (dirs dir dir)
>   nnmh-request-list-1("/home/steve/News/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/nntp+localhost/cache")
>   # bind (dirs dir dir)

The price of lunch says you've got a symlink, ~/News/nntp+localhost,
pointed at `.'

Remove it. :-)

(We went through this last year, when I did something similar under an
old rgnus.  Lars added self-defensive code to scorefile searching at
the time, which is where I ran into this.  Perhaps he could do it here
as well.)

