Rejecting risky extend cells?

[Nick Mathewson]

On Mon, Aug 18, 2008 at 11:54:45PM -0400, Roger Dingledine wrote:
> Hi Nick, others,
> 
> Does this need a proposal, or is it a clearly good patch? Ideas
> suggested by rovv.

Looks good to me.  I'd say check it in, and mark as backportable for
0.2.0.x.

In addition to checking for 2-hop loops, though, we should also check
for 1-hop loops (i.e., extending to yourself).  Do we?

yrs,
-- 
Nick