[openbeosnetteam] Re: Directions [was RE: IRC meeting]

  • From: "David Reid" <dreid@xxxxxxxxxxxx>
  • To: <openbeosnetteam@xxxxxxxxxxxxx>
  • Date: Sat, 19 Jan 2002 00:11:48 -0000

OK, here's an example of what I'm trying to escape.

5.0.4 - net_server

- set 2nd nic to same ip as first (by accident) try to restart the network
- total hang

But hangon you yell, it's a userland stack, how can that be!  Go figure.

- reboot machine, all is well until network starts...
- total hang
- reboot
 etc etc

- remove one nic
- boot
- network fails to start as one card couldn't be found
- click ok on all the boxes network starts
- last error box moans saying that the nic that's left has failed to
initialise or some such

I mean, come on, it's not that hard!  If this is what peopel wnt, then fine,
but personally I'd like to think we could give openbeos a much better
solution.  even if we do write a compatibility layer we'll make sure a lot
of the "nonsense" that was in net_server has been fixed.

david

>
> For my part this is what I would like:
>
> - developping network stack with NewOS
> - developping replacement component for BeOS R5, firstly Network Kit.  For
> the BONE thing, I think we will have something new to redesign when first
> version of OpenBeOS will come out, forking from NewOS.
>
> I know this can be seen as a splitted effort, but ok, at the same time
going
> top to bottom and bottom to top for later merging the results seems to me
a
> good thing...
>
>    Manu
>
> > Also, it's worth noting that there is a plan to have a compatibility
> libray
> > so people can still use net_server calls...
> >
> > The fact of the matter is that unless we move forward we may as well not
> > bother!  This way we move forward but give a nod to the past, which is
> > probably the best of both worlds.
> >
> > david
> >
> > >
> > > From: "Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx>
> > >
> > > > > I think there's space on the openbeos site for the netkit.  I'm
not
> > > > > sure who maintains this.  I'll cc michael phipps on this, maybe
> he'll
> > > let us
> > > > > know. :)
> > > >
> > > > So we are going to reimplement BONE instead of the net=5Fserver=3F
> > > > While I am okay with this, shouldn't discuss this in the OpenBeOS
> > > > mailing list=3F
> > >
> > > Well, this is a moot point, as Travis will be implementing a
kernel-land
> > > stack for NewOS with or without our help. So I guess the kernel-land
one
> > is
> > > the way o go.


Other related posts: