Re: [TLS] Should we require implementations to send alerts?

[Brian Smith <[e-mail]>]

On Fri, Sep 18, 2015 at 4:36 AM, Hubert Kario <[e-mail]> wrote:

> On Friday 18 September 2015 00:58:19 Martin Rex wrote:
> > Easier troubleshooting is IMO a sufficient rationale to justify
> > existence of the alert mechanism and a "SHOULD send the alert before
> > closing the network connection".
> >
> > A "MUST send fatal alert" requirement, however, would be silly (and
> > will be void in face of rfc2119 section 6 anyway).  What would be
> > the semantics of such a requirement anyway?
>
> That's true only if you ignore the situation when TLS 1.4 or TLS 2.0 is
> deployed.
>

So yes, it's no a direct interoperability issue, but it will become one
> in the future.
>

Given a *conformant* TLS 1.3 implementation, that kind of interoperability
problem could only happen if the TLS working group specifically designed it
to happen. In particular, a conformant TLS 1.3 implementation must accept
larger values of ClientHello.client_version.

The same way as TLS protocol version in Client Hello
>

Right. We already have ample evidence that shows it is not reasonable for
TLS 1.3 nor future versions can use ClientHello.client_version to signal
the TLS version, due to broken non-TLS-1.3 implementations. But, unless a
terrible mistake is made, whether or not a conformant TLS 1.3
implementation sends alert will not matter in those interactions.

Cheers,
Brian
-- 
https://briansmith.org/