USB Ethernet test

[James Cameron]

On Tue, Jan 11, 2011 at 02:40:23PM -0300, Martin Abente wrote:
> Researching on the web, I found this [1]. Basically autoip was causing
> confusion when making the users to believe that the ethernet
> connection was _always_ successful.

Caused by conflating two concepts; network interface configured, vs user
perception of a useful network.

> http://www.mail-archive.com/networkmanager-list@gnome.org/msg04835.html

I disagree with Jerone.  The connection is not useless, it is ready.

Once another node joins the network, it will be accessible.  However, I
don't think the connection state should be given to the user as a claim
that they have a useful network that will carry their packets to the
internet.  That's a fundamental design issue.

It can be made to work in our case, just by assigning an IP
automatically, since Sugar can operate well on a network that does not
carry packets to the internet.

p.s. modern consumer DSL and wireless routers in Australia deal with
this issue by adding an indicator; one LED indicates connected to
provider, another LED indicates success of ping to one of a set of
predefined test IPs on the internet.

-- 
James Cameron
http://quozl.linux.org.au/