[TLS] Comments on cached info at TLS meeting

[Stefan Santesson <[e-mail]>]

Recap of comments on cached info during the IETF meeting in Stockholm.

Simon thought that it would be OK to send just the hash as replacement for
the cached data if the client treats the hash as an index that identifies
the replaced data. Doing so avoids guessing by the client. A clarification
should be made in the draft.

Eric made a comment on Jabber that ³now that I see this problem suggested,
this seems pretty gross from an implementation perspective. I think I'd
prefer that the cache hits be returned explicitly in the extension²

I interpret this to mean that Eric wants the server to acknowledge exactly
what data it will replace when it sends its server hello.
This means that we have different opinions on the list whether the server
have to know what data it will replace at server hello exchange.
I¹m personally pretty agnostic about this, but we need to agree on one
approach.

Other than that it was suggested that other objects could be of interest to
define in the draft. We may see proposals on the list.

/Stefan