Re: [TLS] Issue 16: Alert clarifications

[Nelson B Bolyard <nelson at bolyard.com>]

Martin Rex wrote:
> Bodo Moeller wrote:

>> Of course, I totally agree that it makes sense to require fatal alerts
>> before the connection is torn down (except in the close_notify case,
>> obviously), and that it makes a lot of sense for implementations of
>> the current specs to behave like this.  I just don't agree that the
>> current specifcations already require implementations to do this.
> 
> I firmly believe that the SSLv3 spec required it, but a spec weasel
> might try to argue that the TLSv1.0 an later specs have failed to add
> the proper rfc-2119 terminology to various places of the SSLv3 protocol
> description and therefore relaxed the original SSLv3 requirements.

And in fact, one of the top 10 most-used TLS implementations (one of
the newer ones) does not send fatal alerts (at least, not for most
errors), citing just that reason.

I see this becoming a nightmare for people who must diagnose problems.
I'm afraid it's too late for TLS 1.0 and 1.1 already, but I sure would
like to see it fixed in 1.2.

-- 
Nelson B

_______________________________________________
TLS mailing list
TLS at lists.ietf.org
https://www1.ietf.org/mailman/listinfo/tls