On Wed, 25 Oct 2017 14:16:38 +0200 <to...@tuxteam.de> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On Wed, Oct 25, 2017 at 01:06:37PM +0100, Joe wrote: > > On Wed, 25 Oct 2017 07:35:35 -0400 > > Gene Heskett <ghesk...@shentel.net> wrote: > > > > > > > > > > Whereas my theory has always been WRT the search line, that it > > > should first search the /etc/hosts file for a name match, and > > > failing that, query my router, which is running dd-wrt which > > > means its running dnsmasq. > > > > > > > Yes, but that information (among other stuff) lives in > > /etc/nsswitch.conf, not /etc/resolv.conf. > > > > Keep up at the back: why use one configuration file when a couple of > > dozen will easily do the job? > > To be fair, those two do quite different things. The one is the > configuration for the (host name) resolver, whereas the other is > a configuration for generic name services (which encompasses many > other things, like user names, protocol names... you name it). > So nsswitch is "outside" as seen by the resolver.
As I suggested with '(among other stuff)'. > > Of course, one might envision one highly-integrated naming service > doing "all of that", and that was probably Sun's original vision. > But things like the DNS are complex (and important) enough that > they develop a life of their own. The perennial question of whether a piece of information involving A and B should be filed under 'A' or under 'B'... I submit that the name resolution of networked hosts has rather more in common with other network parameters than with, for example, which database might be expected to contain user account passwords. -- Joe
