Was seeing occasional errors with tests run back-to-back.

Also simplify test running.

Also remove usused imports and outdated doc text.

Simplifies launching a Mininet or running tests on one.

Also remove all non-object-oriented legacy Mininet code and update
tests.

User-space compatibility is untested, but most of the code for it is
still in.

Nodes include Switch, Host, and Controller; move these to a separate
file.

This file still could use some attention to hide private functions.
Node seems like a primary class to extend, for adding stuff like Open
vSwitch, so it could benefit from a simpler interface.

Having a file with the same name as its package creates hard-to-diagnose
import error, so rename mininet.py to net.py.  This commit moves utility
functions, generally those dealing with network namespaces and interface
management, into their own file.

Also fix a few minor formatting bugs and clean up logging code (which
had issues when multiple files would import mininet.logging_mod).

Make much of the Mininet code object-oriented around the new Mininet
object, and create a generic way to build a Mininet.

Previously, each network topology was created by Mininet API calls,
which seems much less approachable for new users than passing in a
topology object along with controller, switch, and host classes.

The modified code is all in ripcordtest.py, but will need to replace the
main mininet.py code eventually.

Kernel-space support only, for now.

Forgot to add __init__.py

Use customized StreamHandler from Python logging module to print
only messages for the specified loglevel to the console.

Don't print retry errors by default.

Was causing OSErrors because the process to kill no longer existed.

Much faster now; from 1 s delay to 100 us.

Now, to reference mininet files, use 'import mininet.mininet'.

PYTHONPATH mods are no longer required for installation.

Useful for monitoring the output of a set of nodes.

- invoke ofdatapath with --fail=closed (no controller = no connectivity!)
- stop hosts before switches before the controller
	- this seems slightly more sensible
	- it's the reverse of the startup order
- fix retry() so it kind of works (it's actually helpful I think!)