wrong checksum type for arcfour-hmac-md5

[Luke Howard]

On 15/09/2010, at 5:58 PM, Greg Hudson wrote:

> On Wed, 2010-09-15 at 11:50 -0400, Luke Howard wrote:
>>> But windows doesn't use it...
>> 
>> Hmm, OK, that's a fair argument then.
>> 
>> Greg?
> 
> What's the proposed change, exactly?  Making the mandatory-to-implement
> checksum for RC4 be an unkeyed checksum seems problematic.

That's the proposed change. It doesn't use the simplified crypto profile so it probably won't affect krb5_encrypt/GSS, but it doesn't sound ideal does it.

-- Luke