Re: [TLS] (draft final) ITU Q3/16 Liaison Response

[[e-mail] (Martin Rex)]

Sean Turner wrote:
> Paul Hoffman <[e-mail]> wrote:
>> 
>> These look good. However, it may be premature to say
>>    "renegotiation is being dropped in TLS 1.3"
>> since we are far from finishing TLS 1.3. Maybe change this to
>>     "renegotiation is likely to be dropped in TLS 1.3".
> 
> Yeah that's a good point, we're not done with TLS1.3 so we
> shouldn't presuppose how it's going to turn out.


You probably should also take into account that TLSv1.2 (rfc5246)
"highly recommends" supporting TLS renegotiation:

https://tools.ietf.org/html/rfc5246#appendix-D.4

   D.4.  Implementation Pitfalls

     [...]

   -  Do you support renegotiation, both client and server initiated?
      While renegotiation is an optional feature, supporting it is
      highly recommended.


While you can get away with (and are better off) with _not_ supporting
renegotiation on the server side, you may encounter frequent interop problems
when you do not implement it on the client side -- because it is used
by a lot of servers (at least in TLS).

Some TLS clients even have serious difficulties dealing properly
with a CertificateRequest handshake message in the initial handshake,
because there seem to be more sites requesting client certificates
through renegotiation than in the initial handshake.


-Martin