[TLS] TLS 1.3 and vendor strings?

[Jeffrey Walton <[e-mail]>]

Are there any plans to support free form vendor strings?

The use case is similar to Apple's buggy ECDHE-ECDSA SecureTransport
for OS X 10.8 and iOS 7.

In this case, OpenSSL had to jump through a number of hoops to
identify the potentially affected clients via fingerprinting.
Fingerprinting was not precise, and it potentially captured unaffected
clients when Apple patched at OS X 10.8.4 and iOS 7.0.3. That is, an
OS X 10.8.5 or iOS 7.0.4 client would potentially be identified as
buggy even though it was patched.

Or is there another way to handle the occasional implementation bug like this?

(And to be clear: patching is not always an option. Apple is not like
Microsoft or Linux. Rather, they left a number of hosts unpatched for
the ECDHE-ECDSA bug; and did the same with CVE-2015-1130 (Goto Fail);
and did the same with CVE-2015-1130 (Hidden Backdoor)).