Re: [rtcweb] JSEP fingerprint hash requirements

[Harald Alvestrand <[e-mail]>]

On 10/21/2013 06:38 PM, Martin Thomson wrote:
> On 21 October 2013 07:21, Harald Alvestrand <[e-mail]> wrote:
>> When receiving browser supports both A and B, we could argue that they
>> should be allowed to be different in the name of algorithm agility. But is
>> there a real gain in security achieved by it?
> Those are interesting cases, but they easily solved by saying
> something like "MUST include/implement SHA-256".
>
> I don't think that the hash used by the certificate is actually
> relevant either.  Fingerprints are calculated, not observed or
> extracted.
Welll... if the hash is what I think it is, you need to compute the hash 
of the certificate (in the algorithm specified in the certificate) in 
order to verify the RSA signature of the certificate.

So if you don't understand the hash algorithm the certificate specifies 
.... you can't verify the certificate's signature, and you are open to a 
forged-certificate attack. Again - saying that they have to be the same 
means that you only have to deal with one bad situation (you understand 
neither), and not three bad situations (you are toast because there's 
one of them you don't understand).

But I'm not a REAL crypto expert.  I may have misunderstood something.