A Dimecres 15 Febrer 2006 13:11, David Pashley va escriure: > On Feb 15, 2006 at 11:44, Marco d'Itri praised the llamas by saying: > > On Feb 15, Leopold Palomo Avellaneda <[EMAIL PROTECTED]> wrote: > > > Otherwise, to me, the question is why a program the manages the devices > > > needs when start to create, to check users and groups, when maybe this > > > information is stored in a network device/system, and the udev is the > > > responsible of wake up the interface. I think that this is a dead lock. > > > > Feel free to propose a better design then. > > The only thing I did wonder about is whether udev needs to do a uid/gid > lookup at startup and whether it can be postponed until it is actually > needed. Of course this could already be what it does, as I haven't > checked the code. I did mention to Leopold that I don't feel there is a > bug in udev (although the information added to README.Debian is possibly > wrong).
Yes, David, excuse me if I have done some mistake. You always have thought that it wasn't a bug. I don't know if it's a bug or a bad design or a bad use from me. I just asked because I didn't understand the behavior. > As I see it, there are two separate bugs in my issue. > > 1) nss_ldap does a host lookup even if you use IP addresses in the > server address. maybe your network device is managed by udev, so if you don't have a network interface, what happens? I suppose that loopback network interface works, no? > 2) nss_wins fails badly in the absence of networking. > No 2 is the killer when using udev, wins and ldap. I couldn't test it, but you have always said that it crash. Regards, Leo -- Linux User 152692 Catalonia
pgpEW93TsMp7c.pgp
Description: PGP signature
