On Fri, Oct 15, 2010 at 9:22 PM, David ‘Bombe’ Roden
<[email protected]> wrote:
> On Friday 15 October 2010 22:01:55 Gregory Maxwell wrote:
>
>> JS can be used for a lot of really really nasty tracking and anonymity
>> busting.
>
> So, you trust our Java code but not our JavaScript code?
>
> I disregard the rest of your mail because I get the distinct feeling that you
> are not separating between the “the Freenet web interface” and “arbitrary
> freesites random people insert.”

That is unfortunate, because we've had a simple and easily corrected
communication error.  One which might have been corrected without any
intervention on my part had you simply taken a moment more to read the
rest of my message, but I apologize for being unclear.

I'm not saying much about the trustworthiness of the freenet code.

A browser which has javascript enabled is potentially subject to
executing malicious code from third parties. The question of this risk
existing via freenet is _mostly_ a question of fproxy successfully
detecting and blocking any of the multitude of ways of tricking a
browser into executing code on the page. Or, in other words, the
_browser_ cannot distinguish between the freenet web interface and
arbitrary freesites and so unless fproxy does a heroic job of removing
everything the browser might possibly execute then javascript poses a
significant risk.

The wild continued success of XSS indicates that this is a very hard
problem— browsers try very hard to make "everything work", but that
means that making things not work is tricky.

Also— I used the word mostly above because some JS driven attacks
wouldn't pass through fproxy. E.g. a non-freenet site could use the JS
CSS link-coloration information leak to learn about your use of
freenet if you browser that site with the same browser you use to
access freenet and have JS enabled.
_______________________________________________
Devl mailing list
[email protected]
http://freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to