[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: RFC: Adding MD5-C++ implemetation to kdelibs
From:       Alex Zepeda <jazepeda () pacbell ! net>
Date:       2000-12-12 1:46:29
[Download RAW message or body]

On Mon, Dec 11, 2000 at 03:05:58AM -0500, Dawit Alemayehu wrote:

> Hmmm... correct me if I am wrong here, I have not looked at RFC 2831 so I am short
> on its specific implementation details, but doesn't any/all implementation of Digest
> Authentication require the use of message digests ?  

Yes.

> I am speaking purely of course
> from the prespective of HTTP here so I may be ignorant of the afformentioned RFC.
> If that is the case, then wouldn't we be creating circular dependency here if we say
> integrate MD5 and base64 code into kio and create a separate library for Digest
> Authentication ?  I have a feeling I am missing something.  I am going to look at RFC
> 2831 right away...

No, not really, just an added dependency.  I'm thinking base64+md5 go into
kio, and the digest auth code into another library *or* all three go into
kio.  Since there are valid uses for base64 or md5 code outside of the
digest auth stuff.

> In my tests it never worked if the
> specified algorithm was MD5-sess.  BTW, have you checked in your implementations
> anywhere or is it still in the works ?  I am just curious how we can make this viable to
> all applications that need it.  I am still torn b/n the integration into kio and a standalone
> library.

Hmm.  Well.  I haven't even begun trying to get this to work with HTTP
yet, I'm still wrestling with cyrus-sasl. AHH.  Anyways, the code I've got
does work for the example provided by the SASL/DIGEST-MD5 auth rfc.  I
don't quite have anything to test it on yet (still waiting for
test.smtp.org..).

The main problem is that no e-mail clients today support DIGEST-MD5 auth
over SMTP (or anything else I think), and there are very few HTTP servers
setup with digest auth as well.  Makes for very difficult testing, altho I
wish people would start using these more since they are a step in the
right direction.

- alex

[prev in list] [next in list] [prev in thread] [next in thread]