Dirk-Willem van Gulik wrote:

Unless I missed something we nicely issue a nonce during digest auth
(based on r->request_time) - but when the reply comes in with an
(Proxy-)Authenticate header we use the nonce provided by the client; and
do not check if it was any where near reasonably likely that we issued it.

So I guess

->   The release notes and the digest docs should
        propably contain a warning that we are not
        hardnened against certain replay attacks.

->   Long term we propably want to solve this; e.g.
        by using a hash over a static secret or somethign.

I distinctly remember discussing this at the time - guess I didn't have the energy to fix it then :-)


Cheers,

Ben.

--
http://www.apache-ssl.org/ben.html       http://www.thebunker.net/

"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff

Reply via email to