Re: [Idr] RFC 4684 pedantry - routes with no Route Target

[Jeffrey Haas <[e-mail]>]

On Wed, May 21, 2014 at 06:38:49PM +0200, Robert Raszuk wrote:
> > While a complicating factor certainly, we'd still end up sending along those
> > routes after either end-of-rib for RT-C or the timeout.  (Presuming the
> > implementation follows the SHOULD.)
> 
> After timeout yes, but not after valid set of RTC RT NLRIs followed by
> EOR. .. I would not expect to see those in the latter case.

I believe that's the point of contention.  Should we be applying filters
when there's nothing to match on?

To echo a point Yakov raised in a side chat, what this would mean
protocol-wise would be a NULL RT-C entry.  This is distinct from a default
which is a "match all".

-- Jeff