On Sun, 2008-07-20 at 13:15 +0200, Petter Reinholdtsen wrote: > > Having a hostname of localhost will probably result in problems. If > > you want to look into how it's implemented have a look at > > nslcd/cfg.c:cfg_getdomainname(). What is basically does is look up > > the hostname and uses that to look up the canonical name which > > should be a FQDN. > > I do not understand why you do it like this. It seem to be the wrong > algorithm. I would expect it to only use the DNS config and ignore > the current hostname. If resolv.conf have > > search sub.top.com top.com > > I would expect it to first look for _ldap._tcp.sub.top.com and then > for _ldap._tcp.top.com, independent of the hosts current name and DNS > entry.
Sorry for not replying sooner. I believe that cfg_getdomainname() does more or less the same thing as hostname -d does (at least I don't think I came up with the solution myself). Do you know of a way to reliable access that information (apart from parsing the file yourself)? -- -- arthur - [EMAIL PROTECTED] - http://people.debian.org/~adejong --
signature.asc
Description: This is a digitally signed message part
