This allows the worker to queue up task updates when Flamenco Manager
is unreachable.

Note that asyncio is used for task scheduling & execution, but HTTP connections
and subprocess communication is not (yet). As a result, a blocking POST
call will block proc.stdout.read() in AbstractSubprocessCommand, and
vice versa. This can be solved by using either multithreading or asyncio
to perform those operations.

Instead, use a personal global gitignore file and ignore whatever
files your editor/IDE/OS generates there.

It can register with the Flamenco Manager, but nothing
else. It does save its secret & worker ID in a config
file, though.

http://gulpjs.com

Moving Server Storage to Public
Storage location set automatically

related to issue #35

- a separate configuration file that is ignored in .gitignore files.
  There is a "config.py_tmpl" which can be used as a starting point
  for a custom configuration. If there were not defined a configuration
  file the system will just work using default values and starting
  respective applications on local-host (for developing proposes)

- DATABASE can now be defined inside configuration file and data base
  connection will be established only from starting point of brender
  (brender.py)

- now inside __init__ files is defined module serve() function that is
  used to start application and setup some configuration value. All
  other operations are moved back do controllers.py files. This should
  keep separate application initialisation from application usage.

Organized: organized the dashboard a little
Changes: small type fixes
Changes: fixed bootstrap 3 alert message style
Fixes: Log page fixes and layout fixes
Added: worker_blender_mem and cpu ussage statistics
Added: blender_stats function to worker

We are going to use the Flask framework for both server and client.
This is just a prototype. Actual documentation will follow later.