Re: Contracts and communication channels in Shill [was Re: Shill: capability-based shell]

Hi Anton,

I'm not exactly sure what you mean by through communication channels.

In Shill, there are basically three ways to pass capabilities between
scripts/functions:
1) as function arguments or returns
2) in a capability "wallet" (as an argument or return value)
3) as part of a closure that is passed as an argument or returned

Contracts let you control how capabilities communicated via (1) and (2) are
used (including preventing their use). (3) is trickier. Currently, Shill's
contracts don't give you much control over what capabilities a function
closed over can be used when you invoke it. Of course, you can mitigate
this risk by being careful about what capabilities you give to a script
that might return a closure, and being careful who you give the resulting
closure to. It's also worth noting that all mutable state in Shill is
controlled by capabilities (either for files, etc or for wallets), so on
any given invocation of a function, the authority of its dynamic extent is
limited by the capabilities it received as arguments (including
capabilities closed over by any arguments).

I'm not sure if I addressed your question. Could you elaborate on the kind
of communication pattern you mean?

Cheers,
Scott

On Fri, Oct 10, 2014 at 11:48 AM, Anton Burtsev <[email protected]>
wrote:

> Hi Scott,
>
> I was looking at how Shill uses contracts. I didn't read any of the
> original "contracts" papers, but they look interesting. I'm not sure I
> understand how contracts help me to control the flow of capabilities in
> case when functions actively exchange capabilities through communication
> channels (this is a common case for me). It seems that contracts only
> control function invocations but not communication interfaces, correct?
>
> Thank you,
> Anton
>
>
> On Thu, Sep 25, 2014 at 12:45:58PM -0400, Scott Moore wrote:
> > Hi all,
> > I'm part of the team working on Shill. I'm happy to answer any questions
> > and we're very excited to hear what cap-talk thinks!
> > Cheers,
> > Scott
> >
> > On Thursday, September 25, 2014, Tony Arcieri <[email protected]> wrote:
> >
> > > Apologies if this has been discussed before but It's New To Me(TM), and
> > > relevant to the recent "Shellshock" escapades with bash:
> > >
> > > http://shill.seas.harvard.edu/
> > >
> > > --
> > > Tony Arcieri
> > >
>
> > _______________________________________________
> > cap-talk mailing list
> > [email protected]
> > http://www.eros-os.org/mailman/listinfo/cap-talk
>
>

_______________________________________________
cap-talk mailing list
[email protected]
http://www.eros-os.org/mailman/listinfo/cap-talk