Re: [Last-Call] Benjamin Kaduk's Discuss on draft-crocker-inreply-react-08: (with DISCUSS and COMMENT)

[Kjetil Torgrim Homme <[e-mail]>]

On Fri, 2021-03-05 at 13:52 -0800, Dave Crocker wrote:
> On 3/4/2021 12:41 AM, Kjetil Torgrim Homme wrote:
> >     2.  If R's In-Reply-To: does reference one, then check R's message
> >         content for a part with a "reaction" Content-Disposition header
> >         field, at either the outermost level or as part of a multipart
> >         at the outermost level.
> > 
> > This means a forwarded message will*not*  get its embedded reactions
> > processed.  Well, forwarded messages will typically not have I-R-T set,
> > but if a message includes previous correspondence as an attached MIME
> > document.  
> 
> Not 'typically'.  A forwarded message is not a reply.

I can forward a message as a reply to a different message (drag and
drop it as an attachment), but the "forward" button in my mail reader
will not do that - it will only set References and not I-R-T.

> So I'm not sure how this is a problem.

It's not a problem, just an edge case.

> In any event, if there is no In-reply-to: field, then this specification 
> is not relevant to that message.

There is an In-Reply-To inside the attached message, that's my point.
 It could even be nested further.

> > However, In-reply-to in messages in the attached
> > correspondence will get their reactions processed if they are at the
> > correct relative level in the structure.
> 
> Let's see whether I understand, with an example meant to be more 
> interesting than the one in the specification:
> 
> > From:  me
> > To: you
> > Subject: I just got this message
> > 
> > ---- Forwarded message
> > From: someone else
> > To: me
> > In-Reply-To: a previous message between us
> > Content-Disposition: reaction
> > 
> > U+1F997
> 
> The containing message isn't using MIME, to make the forwarded message 
> an attachment.  I think it doesn't matter, for this example.

Well, a real example would have to use MIME, a MUA should not try to
assign meaning to "---- Forwarded message" or "---- Vidaresendt
melding", but I'll consider your example a shorthand for a proper MIME
structure.

> And the question is whether the contained message's reaction will be 
> processed as a reaction.  The answer depends on how the MUA processes 
> such things.
> 
> In terms of semantics, the reaction is associated with that contained 
> (forwarded) message and not with the upper level (containing) message.

OK, so processing MAY recurse into the MIME structure?  I don't think
the current step 2 allows this.  This can cause duplication in the
presentation of reactions, but I'm OK with this being a quality of
implementation thing (e.g., keep track of messages-ids for messages
with reactions to eliminate the duplication).

-- 
venleg helsing,
Kjetil T.