Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 29 Apr 2015 10:18:02 +0100
From:      Frank Leonhardt <>
Subject:   Re: dig command ?
Message-ID:  <>
In-Reply-To: <>
References:  <> <> <> <>

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help

On 28/04/2015 11:56, Matthew Seaman wrote:
> On 04/28/15 11:34, Arthur Chance wrote:
>> It didn't. What we've got now is a clone of host supplied by the unbound
>> software.
>> No, I've no idea why they didn't clone nslookup and dig as well. :-(
> Because cloning host(1) is easy.  Cloning nslookup(1) -- replicating its
> peculiar behaviour with the reasonably sane tools bundled with
> unbound(8) -- is hard, and not worth it for a program that is in any
> case deprecated.

Actually, nslookup is NOT deprecated. It was deprecated by ISC for a 
period in 2004(!) but was un-deprecated again in the BIND 9.3 
documentation. ISC woke up and realised it was important ten years ago.

As for peculiar behaviour, all UNIX utilities have peculiar behaviour. 
We've had 30 years to get used to what nslookup can and can't do; 
removing it from the base system after 30 years and breaking scripts all 
over the place is a lot worse than being peculiar. And suppose you want 
do debug a resolver issue? dig (and host?) are going to give exactly the 
same results as the local resolver would because they simply call it - 
nslookup won't. Okay, if you're working on resolvers you'll probably 
install the bind-tools port anyway, but that's not the point. The "find" 
command can be pretty odd - should that be removed in favour of 
something that's theoretically more predictable? Or how about dd, which 
is so old its arguments use a syntax that's incompatible with everything 
else since?

BIND (including dig, host and nslookup) was provided as contributed 
software to the base system. I don't see any reason to "clone" it 
(perhaps there is; I just don't see one), so the difficulty or or 
otherwise of creating a work-alike shouldn't matter, should it?

Want to link to this message? Use this URL: <>