tortila as a bad exit

[Roger Dingledine]

On Mon, Aug 11, 2008 at 04:35:22AM -0500, Scott Bennett wrote:
>      The latest consensus and status documents show that the exit node
> nicknamed "tortila" still has the "Bad Exit" flag from only one authority.
> I was wondering what criteria each of the authority operators use to decide
> whether/when to assign the "Bad Exit" flag to a server's listing in their
> authority server's status documents.  Thanks!

(I've been at defcon and black hat the past week, so have only been
peripherally watching the or-talk list.)

My impression from reading the posts is that tortila's craziness is
another instance of bug 779:
https://bugs.torproject.org/flyspray/index.php?do=details&id=779

We're working on fixing it. This looks like a real bug, and a really
bad one at that. Marking all the exit nodes as badexit (which it looks
like we'd have to do by this metric) is going to make it harder to track
it down.

One of the next steps is to try to learn how to induce the bug. We're
working on that from several different angles.

Any help ya'll can provide would be useful. :)

--Roger