Can we get rid of bitmaps?

[MikeBishop]

Personal opinion, plus Andrei's feedback:

I really dislike the bitmap approach -- I'd much rather use the TLS IANA codepoints that exist. It's more extensible, better layered (can just relay uninterpreted data to the crypto components)

• Can we just get this information from the TLS stack?
• Do we need a TLV-style equivalent to the SETTINGS frame?
• Get rid of unsolicited certs and put this back in REQUEST?

This would simplify the SETTINGS value and reconcile it with the algorithm field on the PROOF.

[martinthomson]

The unsolicited certs is the feature that really pushed this in this direction. Those seem to be useful, particularly when it comes to latency.

[MikeBishop]

I'm almost wondering whether there's value in, say, an EXTENDED_SETTINGS extension that provides TLV settings values. The more I try to create extensions in HTTP/2, the more I feel like the fixed 32-bit values are a mite constrained for communicating anything useful. I don't want to build it in this doc (enough frames already!), but a normative reference might simplify the machinery here. If we could just send an array, then use the same codepoints TLS does, we become much more future-proof and can make the negotiation and PROOF frames match. Not as compact, but still....

[grittygrease]

I tend to prefer the EXTENDED_SETTINGS option, it allows the reuse of the TLS IANA registries and should require less maintenance when useful new TLS 1.3 extensions are inevitably introduced.

[martinthomson]

I think that we sorted this one out with the split.