In a recent note, Ulf H{rnhammar said: > Date: Tue, 20 Aug 2002 08:48:43 +0200 > > On Mon, Aug 19, 2002 at 07:27:41PM -0700, Bela Lubkin wrote: > > If there's no user exposure, I don't see why this is any sort of > > security alert at all. If it causes a security problem for servers, > > those servers are still at risk -- people just have to use > > _any other program that does socket I/O_ (including an unpatched Lynx) > > to attack those servers. > I agree with Bela that security of a server should be the responsibility of the server. Any attempt to enforce server security by restrictions on clients ultimately restricts my freedom to program my own computer, to which I have strong philosophical objections.
> Read the second paragraph of Technical Details again. It allows people to > break out of restrictions, which is what security holes are all about. > But Ulf appears to be concerned that this hole may thwart administrators' intent to restrict users to a captive environment, which is a legitimate concern. > telnet and netcat don't handle URL's. Lynx does. > Nonsense. Telnet handles any stream of characters the user cares to type, including the path part of a URL. I've readily used telnet to access WWW servers. This can be as simple as: telnet www 80 GET / (I just tried it; it returned the HTML source of the home page of our server.) -- gil -- StorageTek INFORMATION made POWERFUL ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to [EMAIL PROTECTED]