On Sat, May 30, 2009 at 01:02:19PM +0100, Jason McIntyre wrote:
> On Sat, May 30, 2009 at 11:02:44AM +0100, Federico G. Schwindt wrote:
> > > > > as an argument. can it really do that? even if it can, the quoting
> > > > > still
> > > > > looks incorrect.
> > > >
> > > > yes, it's like that. how the quote will be in the case of:
> > > >
> > > > return-rst(ttl 20)
>
> "return-rst" [ "(" "ttl" number ")" ]
>
> > > >
> > >
> > > so the document is currently incorrect? we need to document this?
> >
> > yes, this is not mentioned and should be added, but the syntax is
> > correct.
> >
>
> ok, diff below. it is a little problematic, since we do not show the
> exact syntax, but adding it would be even uglier.
>
> > > > > second, can return-icmp really take an icmp6code as an argument? is
> > > > > that
> > > > > what the syntax is saying?
> > > >
> > > > yes, it can take a second argument.
> > > >
> > >
> > > yes, but can it take an argument of an icmp*6* code? i know return-icmp
> > > can take an icmp code, and return-icmp6 can take an icmp6 code, but
> > > return-icmp can take a 6 code (or vice versa)? surely that is wrong.
> >
> > yes, there are 3 variants for return-icmp and 2 for return-icmp6.
> > return-icmp and return-icmp6 set both icmp and icmp6 in all of their
> > incarnations. they will apply the defaults if the argument is not
> > passed/needed, while only return-icmp can take both, icmp and icmp6
> > parameters.
> >
>
> i don;t really understand that, but if you're sure it's right, that's
> fine.
>
> this ok then?
> jmc
looks correct, but wait for some of the pf guys for final ok.
f.-
