Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.mira.net.au!news.netspace.net.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!solace!nntp.uio.no!news.maxwell.syr.edu!news.bc.net!unixg.ubc.ca!nntp.cs.ubc.ca!psgrain!news.rain.net!pacifier!deraadt
From: deraadt@theos.com (Theo de Raadt)
Newsgroups: comp.unix.bsd.netbsd.misc
Subject: Re: Why no addusr?
Date: 15 Feb 1997 03:51:32 GMT
Organization: Theo Ports Kernels For Fun And Profit
Lines: 178
Message-ID: <DERAADT.97Feb14205132@zeus.pacifier.com>
References: <none-ya023480001912962244220001@news.infi.net>
<DERAADT.97Feb10191845@zeus.theos.com> <5dtc0g$hd0@cynic.portal.ca>
<1997Feb14.090136@screwem.citi.umich.edu> <5e2sb7$62a@panix2.panix.com>
NNTP-Posting-Host: zeus.theos.com
In-reply-to: tls@panix.com's message of 14 Feb 1997 18:26:31 -0500
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.netbsd.misc:5385
In article <5e2sb7$62a@panix2.panix.com> tls@panix.com (Thor Lancelot Simon) writes:
> As do we. When you have a PR database the size of NetBSD's, however, it's
> easy to get behind. Starting from scratch is a significant advantage here;
> you can do nothing but close PRs for a few weeks, *then* announce your
> neato-keen operating system for public consumption. We unfortunately
> don't have that luxury.
Yes, our PR database is pretty small. That's why we got bored with
fixing our own PR's and started fixing NetBSD and FreeBSD PR's too. A
quick headcount shows that we've fixed about 60 NetBSD PR's that
NetBSD hasn't, and about 25 FreeBSD PR's that FreeBSD hasn't. (You
can check our Changelogs, in general they provide bug details, credit
the person who filed the PR, which camp it was filed to, and the PR
number).
;-)
> Peter Honeyman writes...
> >the fact is, openbsd dominates netbsd, and theo's "bad" attitude has a
> >lot to do with that.
You'd know Peter, you have a "bad" attidude too! ;-)
> I don't understand what you mean by "dominate" and in fact I suspect that
> you're leaving it deliberately vague. Certainly, OpenBSD crows about itself
> quite a bit more than NetBSD does.
Sometimes we do crow a bit (nothing like Linux or FreeBSD ;-).
And sometimes it attracts developers. For example about two weeks ago
someone named Niels Provos contacted me on IRC because he heard
OpenBSD was interested in security problems. I think he heard it
crowed somewhere. And he described a brand new and ingenious source
routing attack which he had described to a couple of people before who
(mostly) had not understood the consequences. In the end this turned
into a security advisory on the BUGTRAQ mailing list. This too is
GOOD publicity. I hope this too will attract other people like Niels
who are interested in working on improving an operating system.
Remember, back in the days when I wanted to make changes to the NetBSD
source tree I was told I couldn't... Within 4 days Niels was working
with someone else on a really neat piece of code inside OpenBSD: a new
format for the passwd field that is significantly much stronger than
the existing DES passwd entries (and can be scaled to be more
computationally intense over time).
That is how we are alive; it is Vibrant. I hate to tell you guys, but
that is also what Linux developent is like.
> I don't think there are any plans to
> change that. Certainly, certain key people involved with OpenBSD make a point
> of slagging NetBSD at almost every opportunity,
^^^^^^^^
Instead your project has developers that put #ifndef __OpenBSD__ into
NetBSD source files so the FREE code, when merged into another system
which happens to #define __OpenBSD__ will experience machine resets,
right? (Hint to those who are curious, go get an alpha locore.S from
January or so.)
And do not come talking to me about slagging and hostility when you
are in a camp with developers who (uncalled for) slag me and OpenBSD
at a large-room cypherpunks meeting.
On the other hand, people who attented Usenix and the OpenBSD BOF know
what my attitudes were towards the `historical issues'. Even core
FreeBSD developers were saying they were happy about the reasonableness
of my attitude!
No, sorry, I think you are making a mistake when you think that
mentioning good things about OpenBSD is automatically a slag of
NetBSD. That must be something you are reading into the message.
> in order to make OpenBSD
> seem active, ubiquitous, and magical.
But most certainly OpenBSD is _very_ active. That's why we have a live
PowerPC port with miniroots and install floppies; that is why we have
new curses fully integrated; that's why we have IPsec cryptographic
network security going into the source tree sometime this week; that's
why we have put significant efforts into getting closer to POSIX. I
don't even need to mention our security changes.
And it is also why we are doing our best to stay abreast of any and
all changes made in NetBSD and FreeBSD, to decide which are good for
us and which are not, to select, incorporate, and possibly even
improve afterwards.
Certainly the list of changes and improvements are far far far too
lengthy to list here. The place to check for new developments is:
http://www.OpenBSD.org/plus.html
> Certainly, certain key people involved with OpenBSD make a point
> of slagging NetBSD at almost every opportunity,
> We generally try to avoid that type
> ov behaviour.
Ah. I guess that is why you are not slagging OpenBSD in this mail.
I am trying not to slag NetBSD in this mail; they can do what they
like. I am simply saying you are wildly inaccurate with your
accusations. And I am slagging you because you are making vague and
false statements about the work of a group of very industrious people.
> We also maintain a rather different philosophy towards changes. We don't
> like halfway or architecturally dubious fixes -- note that I'm not indicating
> that any filed by CITI people necessarily were such, but we do get plenty of
> "fixes" like that, more than a few of which have made it into OpenBSD -- and
By your words OpenBSD is doing numerous halfway or architecturally
dubious fixes. I believe you are ABSOLUTELY MAKING THAT UP and
haven't a leg to stand on.
And why is then that OpenBSD is about 6 errors away from POSIX
conformance now, and NetBSD has none of the 80 or so problems fixed.
But don't trust me, go grab yourself a copy of the NIST POSIX test
suite, and please tell us where Thorsten's work in this area are
`halfway or architecturally dubious fixes'
By the way, if you feel it is to your advantage to bring up the
`i386-specific aha1542 bounce buffer' thread in reply to this I
strongly caution you because numerous OpenBSD people are using 1542's
on their machines. At least one person is using a 64MB PPro200 with a
1542; what a great place to put a scsi tape drive. I myself am
considering doing this (3 scsi busses!) Anyways, it simply isn't
relevant because in the 5 years I've been hacking on BSD kernels only
one other person I know has wanted to do bounce buffering in another
machine (some Arc MIPS machine that Warner is working on...)
> so depending upon what your point of view is, some things won't ever get
> "fixed" in NetBSD, at least not the way the person submitting the PR
> suggested, and some things that are "fixed" in OpenBSD aren't really fixed
> at all.
First of all, I would like discrete examples of which things we have
fixed incorrectly. Certainly to be making such a confident statement
you must have a list somewhere.
Sometimes the PR authors provide wrong fixes (ie. `` I wanted a
solution NOW!! I need to get back to work!!!!'') However quite often
they provide at least a problem description to work from and discover
and implement a fix for.
How about a few very recent PR's and perhaps you'll get the problems
fixed faster.
use a long line counter, netbsd pr#3083, rhialto@polder.ubc.kun.nl
This one was a bug report about mail. The author's suggestion was
right, a simple two line change. Did you fix it yet?
bpf channels are in use; my fix to netbsd pr#3189, Jean-Luc.Richier@imag.fr
This was an interesting kernel bug which was biting the author. The
authors suggestion is not quite right, but it is close to the mark. A
better fix is pretty easy.
do not use PATH, s/gcc/cc/, mouse@Rodents.Montreal.QC.CA netbsd pr#3138
This one pointed out an interesting problem. Did you even look to see
our fix which 4 people were involved in writing and which solves that
and even more problems? By the way, while you are looking at this,
FreeBSD has some improvements in this area too. And keep an eye on
ours too because it has some minor security fixes.
Lastly, I will mention that yesterday Todd Miller fixed 'dd' to work
with 64-bit offsets. I think both FreeBSD and NetBSD will want to
look at that change. That should close at least one PR in each of
your camps.
Thanks Todd, now I no longer need to 'cat /dev/rsd3c > /dev/rsd4c'
to clone disks.
--
This space not left unintentionally unblank. deraadt@theos.com
www.OpenBSD.org -- We're fixing security problems so you can sleep at night.