Re: [TLS] First TLS cached information draft posted

[Simon Josefsson <simon at josefsson.org>]

Min Huang <huangmin123 at huaweisymantec.com> writes:

> Hi,Simon
>
> This "heuristic" determination works well in most scenarioes, 
> but the client still be confused in some specific cases.

Hi.  Yes.  I don't like TLS implementations playing guessing games on
the intended interpretation of received data, though, and that is my
primary concern with this document right now.  I'm wondering how other
implementers feel about this, is it acceptable?

> I think adding a type-specific tag as you mentioned is a doable
> method, and it can solve the problems by now.
>
> And if we will construct a type-specific tag, the new "datasize" 
> field in CachedInformationHash is still necessary? It seems not 
> necessary any more. It can be a policy conformed by the client 
> when caching data or sending a CachedObject.

Right.

/Simon