[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Breach of trust and security (was: Re: sponsor rules)



On Tue, 17 Jul 2001, "Jürgen A. Erhard" wrote:
> I actually don't see what all the fuss is about: a package was
> uploaded, and was buggy.  Big deal.  Happens all the fuckin time.

No, like this it does not. The problem is NOT the fact that buggy stuff was
uploaded -- that indeed happens far more than it should.

Maybe we should have a hall of shame to award bad karma points to people not
responsible enough to do at least minimal tests (i.e. install) the packages
they upload. But that's a flamewar for another time, and I digress.

The problem is that crap from an untrusted source was UPLOADED (and
installed?) into the archive.  I'll not flame the ftp masters if the stuff
really got installed, they're overworked as it stands and I don't recall
"last security bastion" to be in their job description (but do please
correct me if it does). 

But I *will* hold the sponsor at a fault for it.  And the prospective NM for
not even reading the damn docs and running his packages through lintian,
too, I should add (yes, I am one of those that would want his NMs to learn
packaging before applying as packagers).  

The only reason I did not (and will not) bother to try to find out who the
sponsor was, is that we didn't have guidelines (let's not bother about how
wise it is to upload code that runs as root in god-knows-how-many machines
unverified, only because no docs said not to do it), and it is (no matter
how improbable) indeed possible this person actually had a good excuse for
doing it.

An excuse such as an unconscious desire to force sponsoring guidelines to be
written, added to the good will to give us a bit more prime flamewar fuel, a
severe lack of sleep, and not enough remaining brain power/experience to
warn oneself not to upload when sleepy...  but I digress yet again.

> Isn't it called "unstable"?

It's called "unstable", not "root-kit delivery device".  Uploading
unverified (i.e not checked at all), untrusted (i.e. not from a registered
Debian maintainer as far as the Debian project is concerned) stuff that run
as root to the Debian archive is *NOT* a joke.

> PS: This is *very* reminiscent of the old NM flamewar...
Indeed. Guess what? It will not make the problem go away this time either.

Well, I am in need of sleep, or I'd not even bothered to write this ;)
At least it happened to common buggy crap, and not a trojan. Lets not make
little of this, and honour the trouble we went through to verify that
nothing was tampered with the last time we had a security breach.

Mr. Sponsor, do not do it again.  Mr. NM, get out of our sights for 1 month
and learn to package well while at it -- we all will be better off if you
do.  Mr. hmh, go to bed; you're ranting again.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Attachment: pgpOqZ8EHlm80.pgp
Description: PGP signature


Reply to: