Re: [dccp] new draft submitted

[Eddie Kohler <kohler at CS.UCLA.EDU>]

Hi Tom,

> > > Do we need to say what should/must be done with received packets
> > > that have that
> > > service code?  (ignore?  Or is it specified elsewhere?)
> >
> > It's not exactly specified elsewhere, although the right answer is
> > easy to extrapolate: a Bad Service Code DCCP-Reset.  Should this be
> > made explicit?
>
> It depends on how easy you think it is to get the right answer, and  
> how bad
> the consequences of getting the wrong answer are.  I was just looking
> at the spec diffs, noticed that the receive action wasn't talked  
> about here, then did
> a quick search/think for the proper action, and didn't find it.

I think this is pretty easy to figure out, since the Bad Service Code  
reset is described 2 paragraphs earlier, but here's the editorial  
wording change I'll put in.

    o  The value 4294967295 is an invalid Service Code.  Servers MUST
       reject any DCCP-Request with this Service Code value by  
sending a
       DCCP-Reset packet with Reset Code 8, "Bad Service Code".

(used to be: "The value 4294967295 is an invalid Service Code, and  
MUST NOT appear in the Service Code field of a DCCP-Request or DCCP- 
Response packet.")

Eddie


> If you think that someone reading the whole spec would easily know  
> what to do, then
> leave it as it is.  On the other hand, a quick reference might be  
> worthwhile ...
>
> > > The new port registry and service codes language looks good
> > to me.
> > > The new language clears up something I have long thought ambiguous
> > > -- Can different service codes be concurrently listening/connected  
> > > on the same port?  It seems to say "yes" now.
> >
> > Good!
> >
> > Eddie