Hi,

I closed your bug but I also just commented on it saying "if you send
a patch we will apply it" re: disabling timeouts at compile time. I
still maintain that this is a *very bad idea* for servers which will
live on the public internet, but for intranet/firewalled services it
is probably ok or even beneficial to have it as an option.

G

On Sat, Jan 29, 2011 at 6:59 PM, tsuraan <[email protected]> wrote:
>> I think that would still be trivially exploitable by anyone who knew
>> about it.  It seems like that kind of work would just lead to a false
>> sense of security.
>
> I've been thinking about this, and I don't see how it's exploitable,
> so either I didn't express myself well, or I am missing something.  My
> thought is that, in the controller function, I need to get a user's
> authentication cookie in order to track the user's activity.  If the
> user is DOS'ing me by trickling bytes to me, then I will time out
> while trying to read in the cookie, and I will sever the connection.
> If the user gives me a valid authentication cookie and then starts
> trickling more data in, then I can at least log the activities of that
> user, and take action against a person whose identity is known to me.
> Is there a hole in that?
>
>> It might need to be done with a compile-time flag instead of a runtime
>> check for zero.
>
> That would actually be really great.  I basically never want the
> framework's core to do any timeouts.  I've put this in as a feature
> request (https://github.com/snapframework/snap-core/issues/#issue/50)
> so it doesn't get lost.
> _______________________________________________
> Snap mailing list
> [email protected]
> http://mailman-mail5.webfaction.com/listinfo/snap
>



-- 
Gregory Collins <[email protected]>
_______________________________________________
Snap mailing list
[email protected]
http://mailman-mail5.webfaction.com/listinfo/snap

Reply via email to