- From: Daniel W. Connolly <connolly@beach.w3.org>
- Date: Wed, 20 Dec 1995 10:24:36 -0500
- To: Jon Wallis <j.wallis@wlv.ac.uk>
- Cc: BearHeart/Bill Weinman <BearHeart@bearnet.com>, www-html@w3.org
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
In message <m0tSMkY-000oANC@ccug.wlv.ac.uk>, Jon Wallis writes: >At 13:19 19/12/95 -0600, BearHeart/Bill Weinman wrote: >> >>At 10:40 am 12/19/95 -0800, Walter Ian Kaye wrote: >>><A HREF="index.html"><IMG SRC="../gifs/btnhome3.gif" ALT="[Home]" >border=1></A> >>><A HREF="../map.html"><IMG SRC="../gifs/btnmap3.gif" ALT="[Index]" >> >>>(I'm gonna be changing the form and cgi soon, btw, cuz Lynx doesn't like >>>partial URLs -- tho' Netscape handles this form perfectly.) >> >> The problem with the parial URLs may be the "../" references. >> >> Some servers, and perhaps some browsers too, disallow them because >>they've been abused to get around security measures. > >That really shouldn't be a problem if the system is set up right - but since >so many systems are poorly set up in terms of security I can believe it. I think there are two issues that are getting confused here: (1) whether it's OK to use ../../ in an HREF or SRC attribute in an HTML document, (2) whether it's OK to _send_ ../../ in the path field of and HTTP request. (1) is cool, (2) is not. For example, if the example above was fetched from http://www.foo.com/a/b/c.html, then to fetch the [Home] image, the client must combine the value of the HREF attribute with the base URL as per RFC1808, yielding: http://www.foo.com/a/gifs/btnhome3.gif To access the resource at that address, it makes a TCP connection to port 80 of www.foo.com, and sends: GET /a/gifs/btnhome3.gif HTTP/1.0 Accept: image/* What's _not_ cool is to try to sidestep the processing of .. on the client side; that is, to just combine the base and HREF into: http://www.foo.com/a/b/../gifs/btnhome3.gifs (which is _not_ a well-formed HTTP url) and send: GET /a/b/../gifs/btnhome3.gif HTTP/1.0 This is illegal because it is a potential secruity risk. Consider a server whose document root is /usr/local/etc/httpd/docs/ and a client who sends: GET /../../../../etc/passwd HTTP/1.0 Accept: text/plain a naive server implementation might just do: fopen("/usr/local/etc/httpd/docs//../../../../etc/passwd") and give away a bunch of sensitive info. In stead, any server that sees /../ in the HTTP path is supposed to issue a 403 Unauthorized response. (Is this in the HTTP specs somewhere? YIKES! I can't find it in draft-ietf-http-v10-spec-02.txt!!! HTTP-WG folks: this should be addressed in the HTTP 1.0 spec, no? Dan
Received on Wednesday, 20 December 1995 07:41:44 UTC