Latest world NIS woes

[Jacques A. Vidrine]

On Fri, Apr 18, 2003 at 06:28:13PM +0200, Philip Paeps wrote:
> I'm also getting ssh to dump cores on me when I run it as a nis user.  No
> problems as a normal user.  I have 'passwd: nis files' and 'group: nis files'
> in my nsswitch.conf.
> 
> Gdb tells me the following (after compiling ssh with symbols):
> 
> (gdb) bt
> #0  0x2828d1e5 in strlen () from /usr/lib/libc.so.5
> #1  0x282a1ba0 in .curbrk () from /usr/lib/libc.so.5
> #2  0x280a1e5c in pwcopy () from /usr/lib/libssh.so.2
> #3  0x0804cb41 in main (ac=1, av=0xbfbff7c8)
>     at /usr/src/crypto/openssh/ssh.c:261
> #4  0x0804c475 in _start ()
> (gdb) list
> 227             init_rng();
> 228
> 229             /*
> 230              * Save the original real uid.  It will be needed later (uid-swapping
> 231              * may clobber the real uid).
> 232              */
> 233             original_real_uid = getuid();
> 234             original_effective_uid = geteuid();
> 235      
> 236             /*
> (gdb) 
> 
> Looks like it doesn't like the new nss-bits :-)

Huh, probably one of the struct passwd fields is returned set to NULL,
most likely pw_class (which shouldn't be filled-in if you are not
root).

What is the GDB output of `info reg' ?

Cheers,
-- 
Jacques A. Vidrine <nectar at celabo.org>          http://www.celabo.org/
NTT/Verio SME          .     FreeBSD UNIX     .       Heimdal Kerberos
jvidrine at verio.net     .  nectar at FreeBSD.org  .          nectar at kth.se