Complete.Org: Mailing Lists: Archives: freeciv-dev: April 2004:
[Freeciv-Dev] Re: (PR#8457) No assertion in post_receive_packet_player_a
Home

[Freeciv-Dev] Re: (PR#8457) No assertion in post_receive_packet_player_a

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: matusik_s@xxxxx
Subject: [Freeciv-Dev] Re: (PR#8457) No assertion in post_receive_packet_player_attribute_chunk() !!!
From: "Raimar Falke" <i-freeciv-lists@xxxxxxxxxxxxx>
Date: Tue, 6 Apr 2004 07:57:55 -0700
Reply-to: rt@xxxxxxxxxxx

<URL: http://rt.freeciv.org/Ticket/Display.html?id=8457 >

On Mon, Apr 05, 2004 at 09:26:27PM -0700, Jason Short wrote:
> 
> <URL: http://rt.freeciv.org/Ticket/Display.html?id=8457 >
> 
> > [i-freeciv-lists@xxxxxxxxxxxxx - Mon Apr 05 11:05:36 2004]:
> 
> > It is quite possible that there is some way for a client to crash a
> > server. I don't think this is a big enough in practice to do a full
> > solution (audit all input paths, write a random data generator,...).
> 
> What!?!?
> 
> Surely we should audit all input paths?  Any random testing methods 
> should only be an addition to this.
> 
> I know there have been bugs here in the past.  I've tried to write some 
> of them.  Anyone looking for something to work on?

Note that there are two levels here.

The first the checking for a well-formed packet. Is the packet too
short, too long, does it contain too arrays and so on.

The second level is checking for valid values in a well-formed
packet. Does the player tries to move unit which he doesn't own and so
on.

IMHO hard segv can only happen in the first level. We have done a lot
of work in both parts and I'm quite sure that we have a high level of
checking here. But I would be suprised if we overlooked some parts.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
  One nuclear bomb can ruin your whole day.




[Prev in Thread] Current Thread [Next in Thread]