Re: digest-headers-04: which digest is in the trailer?

[Dmitri Tikhonov <[e-mail]>]

Hi Lucas, Roberto,

I do not have a proposal, nor -- Lucas is right here -- do we at
LiteSpeed Tech intend to implement this (or at least not at the
moment).  My question was theoretical, as it looked to me that that
part was missing once I had read the draft.

  - Dmitri.

On Sat, Feb 13, 2021 at 10:41:20AM +0100, Roberto Polli wrote:
> Thanks Dimitri, good catch.
> 
> I agree with Lucas: a new header can work like
> 
> Has-Digest: sha-256, id-sha-256
> 
> If there's consensus it could be reasonable to include it into this spec.
> 
> Wdyt?
> R
> 
> Il lun 8 feb 2021, 23:41 Lucas Pardue <[e-mail]> ha
> scritto:
> 
> > Hey Dmitri,
> >
> > On Mon, Feb 8, 2021 at 10:06 PM Dmitri Tikhonov <
> > [e-mail]> wrote:
> >
> >> Hello,
> >>
> >> In the case where the digest is included in the trailer (Section 10.11
> >> [1]), it may be beneficial for the client to know *which* digest the
> >> server plans to include in the trailer.  This way, the client, too, can
> >> calculate it in the streaming manner.  I wonder whether this use case
> >> has been considered.
> >>
> >
> > The use case probably has been considered but I'm not sure there is a
> > simple solution that can already solve it. If someone knows HTTP-fu and
> > wants to propose a valid way to do it with existing mechanisms, we could
> > incorporate it into the example. Failing that, I think a new header,
> > defined in a standalone document would be path forward.
> >
> > FWIW draft-thomson-http-mice-03 seems like it was part way to a solution
> > [1]
> >
> > Cheers
> > Lucas
> >
> > [1] https://tools.ietf.org/html/draft-thomson-http-mice-03#section-4.1
> >
> >