Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 9 Jul 2012 00:34:34 -0700
From:      Avleen Vig <avleen@gmail.com>
To:        Doug Barton <dougb@freebsd.org>
Cc:        "Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net>, =?ISO-8859-1?Q?Dag=2DErling_Sm=F8rgrav?= <des@des.no>, Garrett Wollman <wollman@bimajority.org>, FreeBSD Hackers <freebsd-hackers@freebsd.org>
Subject:   Re: Replacing BIND with unbound (Was: Re: Pull in upstream before 9.1 code freeze?)
Message-ID:  <CAMjP1K=b8mwqe31m=OqjUV%2BF=B85L4vpfT%2BDj00a1voPB-8TwA@mail.gmail.com>
In-Reply-To: <4FFA7980.4000707@FreeBSD.org>
References:  <CA%2BQLa9B-Dm-=hQCrbEgyfO4sKZ5aG72_PEFF9nLhyoy4GRCGrA@mail.gmail.com> <4FF2E00E.2030502@FreeBSD.org> <86bojxow6x.fsf@ds4.des.no> <89AB703D-E075-4AAC-AC1B-B358CC4E4E7F@lists.zabbadoz.net> <4FF8C3A1.9080805@FreeBSD.org> <20472.51031.308284.775990@hergotha.csail.mit.edu> <4FF8C890.9030408@FreeBSD.org> <CAMjP1KmVDJuKw09UFXb2M6QaL1dD1ocSjMOZLtjKiYFYoF9f4Q@mail.gmail.com> <4FFA7174.7050604@FreeBSD.org> <CAMjP1K=MahXEgHM-gKHFfDpQRDXY_0LGTn0JEE0Zm43%2Bh5jfPA@mail.gmail.com> <4FFA7980.4000707@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Jul 8, 2012 at 11:26 PM, Doug Barton <dougb@freebsd.org> wrote:
> On 07/08/2012 23:16, Avleen Vig wrote:
>> On Sun, Jul 8, 2012 at 10:51 PM, Doug Barton <dougb@freebsd.org> wrote:
>>> On 07/08/2012 22:43, Avleen Vig wrote:
>>>> It would be silly not to keep bind-tools in base.
>>>
>>> Sounds easy, but not so much in practice. Keeping any of the code
>>> doesn't solve the problem of the release cycles not syncing up. And for
>>> the vast majority of users needs the tools we will import will be more
>>> than adequate.
>>
>> The question I keep asking myself is:
>>   "Is this best for the users?"
>
> Carrying BIND code in the base that is past EOL is not good for the
> users, period. Everything else we're discussing is an implementation
> detail.

I think the "everything else we're discussing is an implementation
detail" is the part we'll have a problem with.
Although Garrett's reply  to my email makes sense too.

>> Linux has `nscd` which is a nice caching resolver, but most
>> distributions still carry bind-tools in the default install.
>
> A) You're wrong about "most." and B) The Linux distros have a default
> set of packages. There is no "base" like there is in FreeBSD. (Thus,
> your analogy is flawed.)

That's not *really* true, there is a "base" like FreeBSD, but what we
consider core userland tools like `ls`, come in a package (coreutils).

> That said, I still believe that our idea of what should, and should not
> be, in the base system is seriously flawed, and needs to be completely
> redone. But that's never going to happen, so I'm trying to work with
> what we've got.

Agreed. The idea of a "minimally functional system" itself might be
flawed. Do you consider having `dig` and `host` essential in a
minimally functioning system? I do.
It's pretty f'king hard to resolve problems with installing the
bind-utils port, if you don't know how to test your DNS :-)

The issue is also one of barrier-to-entry. By removing `dig` and
`host`, I think we're making things unnecessarily more difficult for
people who don't *know* FreeBSD. `dig` and `host` a universally
standard tools for doing DNS lookups. Taking them away in base to
replace them with something else just seems like something that won't
really *help* users.

Yes, I'm going to be a stickler and say that having EOL code in base
isn't the end of the world. It's not ideal, but really.. what is it
breaking?
If there's a security vulnerability, sure, I understand that it might
suck without support from ISC to patch dig/host/nslookup, but when was
the last time that happened?



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAMjP1K=b8mwqe31m=OqjUV%2BF=B85L4vpfT%2BDj00a1voPB-8TwA>