Implement Connection.__del__, which is almost certainly going to trigger
more bugs down the line, because the state of the Connection instance is
not guranteed during __del__. Meanwhile, it is temporarily needed for
deployed-today Ansibles that have a buggy synchronize action that does
not call Connection.close().

A better approach to this would be to virtualize the guts of Connection,
and move its management to one central place where we can guarantee
resource destruction happens reliably, but that may entail another
Ansible monkey-patch to give us such a reliable hook.

Part of an effort to make resource management a little more explicit.

When a Broker() is running with install_watcher=True, arrange for only
one watcher thread to exist for each target thread, and to reset the
mapping of watchers to targets after process fork.

This is probably the last change I want to make to the watcher feature
before deciding to rip it out, it may be more trouble than it is worth.

Full output of failed test

```
ERROR: test_okay (__main__.FakeSshTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "tests/ssh_test.py", line 16, in test_okay
    ssh_path=testlib.data_path('fakessh.py'),
  File "/home/alex/src/mitogen/mitogen/master.py", line 650, in ssh
    return self.connect('ssh', **kwargs)
  File "/home/alex/src/mitogen/mitogen/parent.py", line 463, in connect
    return self._connect(context_id, klass, name=name, **kwargs)
  File "/home/alex/src/mitogen/mitogen/parent.py", line 449, in _connect
    stream.connect()
  File "/home/alex/src/mitogen/mitogen/ssh.py", line 104, in connect
    super(Stream, self).connect()
  File "/home/alex/src/mitogen/mitogen/parent.py", line 395, in connect
    self._connect_bootstrap()
  File "/home/alex/src/mitogen/mitogen/ssh.py", line 116, in
_connect_bootstrap
    time.time() + 10.0):
  File "/home/alex/src/mitogen/mitogen/parent.py", line 207, in
iter_read
    (''.join(bits)[-300:],)
mitogen.core.StreamError: EOF on stream; last 300 bytes received:
'Usage: fakessh.py [options]\n\nfakessh.py: error: no such option: -o\n'
```

I could not get Python 2.5 or earlier to work. Too many packages
(critically docker) don't support it.

This eliminates the possibility of the filesystem and setup.py
diverging, as had happened with ansible_mitogen/connection/ vs
ansible_mitogen/connection.py

Related to issue #141.

Now there is a single deadline calculated by the parent.Stream
constructor, and reused for both SSH and sudo.

Was triggering a crash indirectly due to Ansible passing us Unicode
strings. Needs a better fix.

SSH command size: 439 (+4 bytes)
Preamble size: 8941 (no change)

This _increases_ the size of the first stage, but
- Eliminates one of the two remaining uses of `sys`
- Reads the preamble as a byte-string, no call `.encode()`
   is needed on Python 3 before calling `_()`

SSH command size: 435 (-4 bytes)
Preamble size: 8962 (no change)

os.execl is the same as os.execv, but it take a variable number of
arguments instead of a single sequence.

SSH command size: 439 (-4 bytes)
Preamble size: 8962 (no change)

SSH command size: 443 (-5 bytes)
Preamble size: 8962

SSH command size: 448 (-5 bytes)
Preamble size: 8941 (no change)

NB: The 'zip' alias was absent in Python 3.x, until Python 3.4. This
should change be reverted if Python 3.0, 3.2, or 3.3 support is
required.

SSH command size: 453 (no change)
Preamble size: 8941 (-5 bytes)

SSH command size: 453 (-8 bytes)
Preamble size: 8946 (no change)

SSH command size: 461 (-8 bytes)
Preamble size: 8946 (no change)

Since python has reached the last statement this should occur anyway.

SSH command size: 469 (-11 bytes)
Preamble size: 8946 (no change)

Although the source is longer, the _compressed_ length is reduced.

SSH command size: 481 (down 1)
Preamble size: 8946 (no change)

SSH command size: 482 bytes (no change)
Preamble size: 8946 bytes (down 33)

Not strictly a rival, but has enough commonalities to be worth noting

All outgoing links checked with

```bash
cd docs
make linkcheck
```

```bash
cd docs
make
```

to run a webserver that automatically rerenders whenever the rST is
modified.

A first small mea culpa to all my testing sins of late :)

This actually addresses multiple problems:

* Single-file programs were broken, since the fix introduced in
  6931cc10c4a1e08161a58cf4b6b555b71cdc62b5 caused builtin_find_module()
  to start indicating __main__ can always be loaded locally. That's
  broken, and there might be more cases where the same problem will crop
  up.

  Since it was indicated __main__ could be loaded locally, the built-in
  import machinery was allowed to attempt that (since we remove __main__
  from sys.modules during bootstrap), which caused a safety check to
  fire in the bowels of Python:

      "Cannot re-init internal module %.200s"

* The check for presence of the whitelist was totally broken, since the
  whitelist is never an empty list. Therefore 'self' was being returned
  for every module, including extension modules like 'termios'.

I have hand-verified this does not break the fix for issue #113. I
looked at writing a test for that, but it requires a Docker container
(or similar) with an ancient version of Ansible installed. Will open a
separate ticket tracking this.