mailing list archives
Re: Mathematica on Linux /tmp/MathLink vulnerability
From: Marsh Ray <marsh () extendedsubset com>
Date: Sat, 15 May 2010 13:15:09 -0500
On 5/14/2010 6:19 AM, paul.szabo () sydney edu au wrote:
Personal opinions (hoping not to start a flame war) on your questions:
A. Does anyone think there would be much gained by me requesting (or
insisting on) a CVE number?
I do not see the need for CVEs for such "trivial" (easily verifiable)
It's becoming apparent that's trivial to some is deep hacking to others.
I find them useful when problems or fixes are reported in
"obscure" software (e.g. in binary-only proprietary distributions), so
as to identify same-known-problem reports.
Maybe for multiple-affected vendor situations.
B. Does anybody actually care about local escalations any more, or is
everyone just expected to have their very own personal virtual private
cloud for a security boundary?
We each must protect our own cocoons.
It's easier if we can agree on a set of expectations and stick by them,
particularly for primitives like file permissions and process isolation.
If there's little agreement for a particular thing, I might want to
build that cocoon a little differently. Which is why I asked the question.
Saving the world is impossible.
That's not been proven.
Full-Disclosure - We believe in it.
Hosted and sponsored by Secunia - http://secunia.com/