RE: Failed assertion in sockbuf.c

["Mark de Jong" <dejongm@secdog.com>]

Is the 'passwd' command known to work under Solaris 8 with Openldap? I
found this in my search for clues.

"Solaris' "passwd" command is hard-coded to work with local UNIX files,
NIS or
NIS+ - and that's it;  getting it to work with anything else at all can
be a
right royal PITA.  This is documented as a Solaris bug, but as yet isn't
fixed.  In the interm, try "passwd -r files" to change local user
passwords."

Granted that quote is two years old but does it still stand? Would it be
safe to completely remove the Solaris ldap libs and start from scratch
or does that cause a dependancy issue?


-----Original Message-----
From: owner-openldap-software@OpenLDAP.org
[mailto:owner-openldap-software@OpenLDAP.org] On Behalf Of Howard Chu
Sent: Friday, January 11, 2002 3:08 PM
To: Clint Adams
Cc: Mark de Jong; OpenLDAP-software@OpenLDAP.org
Subject: RE: Failed assertion in sockbuf.c


The truss shows something that may be cause for concern; the process in
question (passwd) has wound up loading both Solaris' bundled LDAP
library and the OpenLDAP library. It is very likely that these libraries
are interfering with each other. Still, this cannot be confirmed without
actually seeing things under a debugger. As long as you build the
OpenLDAP libraries with debugging info, and you have root privilege, you
can execute the passwd command using the debugger and when it hits the
assertion you can get a stack trace.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support

> -----Original Message-----
> From: Clint Adams [mailto:clint@gcfm.net]
> Sent: Friday, January 11, 2002 11:45 AM
> To: Howard Chu
> Cc: Mark de Jong; OpenLDAP-software@OpenLDAP.org
> Subject: Re: Failed assertion in sockbuf.c
>
>
> > The assert message is not really useful without a stack trace. Do 
> > you have any way of debugging the binaries to obtain this?
>
> I can't do a stack trace, but this is a truss from the password that 
> gives me that error.
>