[or-cvs] [tor/master] LetsKillNoConnect removes support for .noconnect This is a patch to remove support for .noconnect. We are removing .noconnect because of a talk at Defcon 17 by Gregory Fleischer.

[Robert Hogan]

On Monday 10 August 2009 03:14:53 Nick Mathewson wrote:
> Author: Jacob Appelbaum <jacob at appelbaum.net>
> Date: Sat, 8 Aug 2009 19:15:22 -0700
> Subject: LetsKillNoConnect removes support for .noconnect
> Commit: 33762b529694f58d7640ebcbef1bc0f940419c89
>

Um, why? Gregory's talk introduces it as a more reliable way of 
establishing if a browser is using Tor. Isn't that what the DNSEL already 
does? If it's more reliable than DNSEL alone, why not offer his javascript 
to sites as a method of corroborating DNSEL results? 

I agree the discovery of an unintended effect is a good reason for removing 
a feature temporarily, just in case, but would it not be useful to 
reinstate it as '.isusingtor' and allow remote sites to make use of it?