There is no need for any client to call tool directly.

This results with calling correct methods on the server.

The module is to access Distributor not the whole tool.

This results with calling correct methods on the server.

- some method names were changed to make them clearer
- new intermediate class 'SlapOSTester'

- method waitInstance moved from runner to communicator
- new simple methods in communicator for request and wait instance
- new methods to modularize runner code
- communicator initialization code moved in runner

Variable name changed from jsonData to json_data.
Old comment deleted.

Some methods in the slapos master communicator no longer exist. The unit test for scalability was updated accordingly.

An update in zope.testbrowser dropped the mechanize dependency used by the browser. Now it uses WebTest instead.

Details of the update: https://pypi.python.org/pypi/zope.testbrowser/5.2.1

- Most of the communication with slapos master is done using the new Slapos Master Communicator instead of Slapos Controller:
  * supply, request, check instance status, check software status.
- Refactor of related methods (waitInstance, isSoftwareReleaseReady, etc.)
- Initialization of communicator: initialize connection to SlapOS usign key and certificate
- General refactor and cleanup: some unnecesary sentences, comments and methods were removed.

- The base of this code comes from some work done by Rafael in here:
    https://lab.nexedi.com/nexedi/slapos.toolbox/blob/master/slapos/agent/tester.py
- Refactor in communication with slap objects for supply and request.
- Added method to check if instance is requested.
- Refactor in log calls.

In order to safely use non-XML allowed characters coming from various
sources wrap basestring into xmlrpclib.Binary before sending it to the
server.

Firefox and Xvfb aren not provided by the tested SR,
but rather by a separate SR that is installed by the testnode.

Also if process_timeout is available in received configuration setup
process_manager with transmitted value.

Backward compatibility is assured, that if no process_timeout is
provided, nothing is being updated.

Also process_timeout not set on master (the default), will not change
local default.

Since a process might exit by himself at any time, handle exceptions when
we try kill processes

This prepares for the removal of --firefox_bin and --xvfb_bin options. The
API between the testnode and the tested software must be exclusively generic.

Conflicts:
	product/ERP5Type/tests/ERP5TypeFunctionalTestCase.py

/reviewed-on nexedi/erp5!505

/reviewed-on nexedi/erp5!346

This caused two problems:
 - If Xfvb was killed and left a stale lock file, we don't reuse this
   display. After some time we end up having "All display locked"
   exception but no Xvfb process running.
 - Xvfb's slapos is patched to store control sockets and lockfiles in
   $TMPDIR ( nexedi/slapos@58462c24 ).
   We don't have $TMPDIR set in testnodes, but this would be a problem
   if we ever set it.

Because Xvfb can detect a stale lock file upon startup and start anyway,
the easy way running Xvfb for each display and see if it's still running
after a few seconds.