[Techtalk] Theory vs. practice

Julie jockgrrl at austin.rr.com
Tue Jan 15 21:22:43 EST 2002


> 
> On Mon, Jan 14, 2002 at 02:15:38AM -0600, Julie wrote:
> > What I'm describing has nothing at all to do with coding security.  It
> > has everything to do with very basic (as in, first year university)
> > skills -- things like, oh, never copy more data into a buffer than it
> > can handle, validate your parameters, know what your code is actually
> > supposed to be doing.
> 
> Ho hum, this is my third post in this thread along these lines, but some
> universities never teach these skills. For starters you can get a whole
> degree in some places using languages where you don't do memory
> allocation and never have to deal with overflowing buffers.

Yes, I certainly do understand that.  But again, these are such
basic skills that I can't believe people see them as anything
special.

Before people figured out how to do buffer overflows (which I
think we can blame on Robert "Wormer" Morris, Jr. ...) those
sorts of bugs were just "bugs".  You got a core dump, and then
you started over.

What I think of as "security bugs" is code like this --

	if (stat (file, blah) == 0 && blah.st_uid == me)
		chown (file, me, new_group);

where the fact that there is a race between the stat and the
chown exists.  It's more like "unintended consequences of
properly written code" than "exploitable incorrectly written
code".
-- 
Julianne Frances Haugh             Life is either a daring adventure
jockgrrl at austin.rr.com                 or nothing at all.
					    -- Helen Keller



More information about the Techtalk mailing list