[Date Prev][Date Next]
Re: slapd coredumps on search: use of assertions? (ITS#1308)
- To: openldap-devel@OpenLDAP.org
- Subject: Re: slapd coredumps on search: use of assertions? (ITS#1308)
- From: Peter W <email@example.com>
- Date: Sun, 9 Sep 2001 11:45:10 -0400
- Content-disposition: inline
- In-reply-to: <200109091534.f89FYvc57678@galois.openldap.org>; from firstname.lastname@example.org on Sun, Sep 09, 2001 at 03:34:57PM +0000
- References: <200109091534.f89FYvc57678@galois.openldap.org>
- User-agent: Mutt/1.2.5i
On Sun, Sep 09, 2001 at 03:34:57PM +0000, email@example.com wrote:
> I have tried it with cvs REL_ENG_2 version and it also segfaults. The only
> difference is that 2.0.12 always segfaults but REL_ENG_2 segfaults only if
> this search is the first search.
> slapd: acl.c:101: access_allowed: Assertion `attr != ((void *)0)' failed.
Isn't there something better to use than generic assertion calls?
Presumably the assertion in this case is checking something worthwhile,
but it looks to me like a failed assertion usually (always?) makes the
process die and dump core. Dumping core to me always suggests possible
security issues. Would it be safer and cleaner to have some function like
openldap_warn_and_die() to be used instead of assertions?
> On Wed, 5 Sep 2001 Kurt@OpenLDAP.org wrote:
> > I note that I committed a fix to filter indexing code to
> > HEAD and OPENLDAP_REL_ENG_2 recently which may resolve
> > this issue. I suggest taking a look at OPENLDAP_REL_ENG_2
> > (available via AnonCVS).
Happy one billion!!!