WG: [Xorp-hackers] Bugzilla #161

'Kristian Larsson' kristian@juniks.net
Thu, 8 Sep 2005 17:00:38 +0200


On Thu, Sep 08, 2005 at 01:41:01PM +0200, "Patrick Preuß" wrote:
> > --- Ursprüngliche Nachricht ---
> > Von: 'Kristian Larsson' <kristian@juniks.net>
> > An: Patrick Preuss <deathdealer@gmx.net>
> > Kopie: xorp-hackers@xorp.org
> > Betreff: Re: WG: [Xorp-hackers] Bugzilla #161
> > Datum: Wed, 7 Sep 2005 13:36:31 +0200
> > 
> > On Wed, Sep 07, 2005 at 12:05:35PM +0200, Patrick Preuss wrote:
> > > Hello Kristian,
> > > 
> > > > Sorry for dual posts here but I couldn't confirm
> > > > this before as I did not have access to a xorp box
> > > > When doing show route table ipv4 unicast final I
> > > > see:
> > > > Xorp> show route table ipv4 unicast final 
> > > > Network 10.0.0.0/24
> > > >   Nexthop := 192.168.77.1
> > > >   Metric := 0    Protocol := static Interface := eth0 Vif := eth0
> > > > Network 192.168.77.0/24
> > > >   Nexthop := 192.168.77.90
> > > >   Metric := 0    Protocol := connected Interface := eth0 Vif := eth0
> > > > 
> > > > Isn't that information enough?
> > > 
> > > Have used <TAB> or <?> for completing the line, I know there is a
> > problem
> > > with both in the current cvs, I've found the point for <?> completion so
> > far
> > > It is in the cli_command.cc there are two entrys with <Enter> for the
> > <?> I
> > > have found the formatting line, for the <Tab> not jet.
> > 
> > <?> gives me an extra blank line
> >  Xorp> show route table ipv4 unicast final ?
> >  Possible completions:
> >   <[Enter]>       Execute this command
> > 
> >   brief           Show IPv4 winning routes
> >   |               Pipe through a command
> > 
> > while <tab> is missing one :
> >  Xorp> show route table ipv4 unicast final  
> >  `final' is ambiguous.
> >  Possible completions:
> >   <[Enter]>       Execute this command  brief Show IPv4 winning routes
> >   |               Pipe through a command
> > 
> > It seems that the missing blank line exist 
> > on every node that are both executable and 
> > have sub-nodes.
> > 
> > Patrick, I checked the code regarding administrative
> > distance and protocol. I now see what you mean and I
> > agree that it's not a pretty solution. There should be
> > something like origin: ospf.0
> > where ospf.0 would be your first ospf process. Later
> > when we (cause we will :) ) support several routing 
> > processes it would just use a different number. 
> > Just my $0.02
> > 
> > Regards,
> > Kristian
> > 
> 
> Hello Kristian,
> 
> can you try the attactched cli_command.cc should go the dir cli. i think no 
> it should work fine.
Works!
Good job. Should be included in the CVS!

//Kristian