Ajax one panel

[info at secappdev.org (Johan Peeters)]

thanks for the clarification, Gary. So I would like to restate my 
question as follows: do you know of any (atempts at) implementations of 
a sandbox based on closures? It seems to me that there are some tricky 
issues like how the permissions get into the closure or how privileged 
code is called. But that's probably due to the limitations on my 
imagination.

Meanwhile, at Artima 
(http://www.artima.com/weblogs/viewpost.jsp?thread=161019) and 
ObjectMentor 
(http://blogs.objectmentor.com/ArticleS.MichaelFeathers.ItsTimeToDeprecateFinal), 
people are also discussing the 'final kludge', calling for 'final' to 
become deprecated. It would be nice to think there is something that 
could take its place :-)

kr,

Yo

Gary McGraw wrote:
> You guys are both right, but talking about different things.
>
> You can kludge pretend closure for a security critical section using the static final nonsense to make a hole on the 
> unbound environment.  I described this in some detail in a 1999 javaworld article.  Havung closure would lead to a 
> better sandbox.
>
> The luring attacks that john describes were not what I was getting at, but they lend more weight to the argument for 
> closure.
>
> Yay, language arcana!
>
> gem
>  
>
>  -----Original Message-----
> From:         Johan Peeters [mailto:info at secappdev.org]
> Sent: Sun May 21 09:08:14 2006
> To:   John Steven
> Cc:   Gary McGraw; Mailing List, Secure Coding; SSG
> Subject:      Re: [SC-L] Ajax one panel
>
> We may be at cross purposes. I understand your concern about luring 
> attacks, John. I am sure you are right and they are feasible, but I 
> interpreted Gary's comment as meaning 'closures can be used to build a 
> more reliable sandbox'. But maybe this is not what is meant?
>
> kr,
>
> Yo
>
> John Steven wrote:
>
> > Johan,
> >
> > Yes, the attacks are feasible. Please refer to the Java language  spec. 
> > on inner/outer class semantics and fool around with simple test  cases 
> > (and javap -c) to show yourself what's happening during the  compile step.
> >
> > Attacks require getting code inside the victim VM but mine pass  
> > verification silently (even with verifier turned on). Calling the  
> > privileged class to lure it into doing your bidding requires only an  
> > open package (not signed and sealed -- again see spec.) and other fun- 
> > and-excitement can be had if the Developer hasn't been careful enough  
> > to define the PriviledgedAction subclass as an explicit top-level  class 
> > and they've passed information to-and-fro using the inner class  
> > syntactic sugar--rather than explicit method calls defined pre- compile 
> > time.
> >
> > ----
> > John Steven
> > Technical Director; Principal, Software Security Group
> > Direct: (703) 404-5726 Cell: (703) 727-4034
> > Key fingerprint = 4772 F7F3 1019 4668 62AD  94B0 AE7F
> > http://www.cigital.com
> > Software Confidence. Achieved.
> >
> >
> > On May 21, 2006, at 8:23 AM, Johan Peeters wrote:
> >
> >
> > > That sounds like a very exciting idea, but I am not sure about the  
> > > mechanics of getting that to work. I assume the permissions for the  
> > > untrusted code would be in the closure's environment. Who would put  
> > > them there? How would the untrusted code call privileged code?
> > > Has anyone done this?
> > >
> > > kr,
> > >
> > > Yo
> > >
> > > Gary McGraw wrote:
> > >
> > >
> > > > Hi yo!
> > > > Closure is very helpful when doing things like crossing trust  
> > > > boundaries.  If you look at the craziness involved in properly  
> > > > invoking the doprivilege() stuff in java2, the need for closure is  
> > > > strikingly obvious.
> > > > However, closure itself is not as important as type safety is.    So 
> > > > the fact that javascript may (or may not) have closure fails in  
> > > > comparison to the fact that it is not type safe.
> > > > Ajax is a disaster from a security perspective.
> > > > gem
> > > > -----Original Message-----
> > > > From:     Johan Peeters [mailto:info at secappdev.org]
> > > > Sent:    Sat May 20 15:44:46 2006
> > > > To:    Gary McGraw
> > > > Cc:    Mailing List, Secure Coding; SSG
> > > > Subject:    Re: [SC-L] Ajax one panel
> > > > I think Java would have been a better language with closures, but  I 
> > > > am intrigued that you raise them here. Do you think closures  present 
> > > > security benefits? Or is this a veiled reference to Ajax?  I guess 
> > > > JavaScript has closures.
> > > > kr,
> > > > Yo
> > > > Gary McGraw wrote:
> > > >
> > > >
> > > > > Ok...it was java one.  But it seemed like ajax one on the show  
> > > > > floor.   I participated in a panel yesterday with superstar bill  
> > > > > joy.  I had a chance to talk to bill for a while after the gig  and 
> > > > > asked him why java did not have closure.  Bill said he was on  a 
> > > > > committee of five, and got out-voted 2 to 3 on that one (and  some 
> > > > > other stuff too).  You know the other pro vote had to be guy  
> > > > > steele.  Most interesting.  Tyranny of the majority even in java.
> > > > >
> > > > > Btw, bill also said they tried twice to build an OS on java and  
> > > > > failed both times.  We both agree that a type safe OS will happen  
> > > > > one day.
> > > > >
> > > > > Here's a blog entry from john waters that describes the panel  from 
> > > > > his point of view.
> > > > >
> > > > > http://www.adtmag.com/blogs/blog.aspx?a=18564
> > > > >
> > > > > gem
> > > > > www.cigital.com
> > > > > www.swsec.com
> > > > >
> > > > >
> > > > > Sent from my treo.
> > > > >
> > > > >
> > > > > -------------------------------------------------------------------- 
> > > > > --------
> > > > > This electronic message transmission contains information that  may be
> > > > > confidential or privileged.  The information contained herein is  
> > > > > intended
> > > > > solely for the recipient and use by any other party is not  
> > > > > authorized.  If
> > > > > you are not the intended recipient (or otherwise authorized to  
> > > > > receive this
> > > > > message by the intended recipient), any disclosure, copying,  
> > > > > distribution or
> > > > > use of the contents of the information is prohibited.  If you  have 
> > > > > received
> > > > > this electronic message transmission in error, please contact the  
> > > > > sender by
> > > > > reply email and delete all copies of this message.  Cigital, Inc.  
> > > > > accepts no
> > > > > responsibility for any loss or damage resulting directly or  
> > > > > indirectly from
> > > > > the use of this email or its contents.
> > > > > Thank You.
> > > > > -------------------------------------------------------------------- 
> > > > > --------
> > > > >
> > > > > _______________________________________________
> > > > > Secure Coding mailing list (SC-L)
> > > > > SC-L at securecoding.org
> > > > > List information, subscriptions, etc - http://krvw.com/mailman/ 
> > > > > listinfo/sc-l
> > > > > List charter available at - http://www.securecoding.org/list/ 
> > > > > charter.php
> > >
> > > -- 
> > > Johan Peeters
> > > program director
> > > http://www.secappdev.org
> > > +32 16 649000
> >
> >
> >
> >
> > ---------------------------------------------------------------------------- 
> >
> > This electronic message transmission contains information that may be
> > confidential or privileged.  The information contained herein is intended
> > solely for the recipient and use by any other party is not authorized.  If
> > you are not the intended recipient (or otherwise authorized to receive this
> > message by the intended recipient), any disclosure, copying, 
> > distribution or
> > use of the contents of the information is prohibited.  If you have received
> > this electronic message transmission in error, please contact the sender by
> > reply email and delete all copies of this message.  Cigital, Inc. 
> > accepts no
> > responsibility for any loss or damage resulting directly or indirectly from
> > the use of this email or its contents.
> > Thank You.
> > ---------------------------------------------------------------------------- 

-- 
Johan Peeters
program director
http://www.secappdev.org
+32 16 649000