Python Packaging Setup

[kb1lqc]

Starting off by packaging the project from the start. Since we're starting with python 3 I'm using the following guides:

• http://www.diveintopython3.net/packaging.html

• https://packaging.python.org/tutorials/distributing-packages/

• https://docs.python.org/3/library/distribution.html

• Ensure the package is built correctly (installed python modules), there is no program to run yet. However, unit testing should accomplish this.

• Automate package pushing to the pypi servers

[kb1lqc]

Cool, we can create windows installation programs!

[kb1lqc]

Oh wow we can automate the python package deployment! https://docs.travis-ci.com/user/deployment/pypi/

[kb1lqc]

I can see that my locally generated source distribution can install and be imported with from faradayio import faraday

>>> from faradayio import faraday
>>> dir(faraday)
['Faraday', 'Monitor', 'TunnelServer', '__builtins__', '__cached__', '__doc__', '__file__', '__loader__', '__name__', '__package__', '__spec__', 'pytun', 'sliplib', 'threading']

This appears to have the proper Faraday, Monitor, and TunnelServer classes available as expected!

[kb1lqc]

I have Travis-CI encryption working on my computer after installing ruby. However. I'm concerned that I need to perform the encrypt command when working in a faradayrf checked out repository. This may work if I encrypt it with my personal kb1lqc/faradayio repository but when we want to upload it from the FaradayRF repo it may fail to decode since Travis-CI associates each repo with a different public/private key.

Is this true @reillyeon @hdkmike @el-iso @lqdev?

[kb1lqc]

We will see how this goes. I manually made travis encrypt use faradayrf/faradayio as the repository so hopefully it signed with that private key.

[kb1lqc]

Closing since the main intent of this ticket has been achieved and #32 covers actual deployments from Travis CI.

I have performed an automated upload attempt on build#86. It appears that the deployment was attempted but halted since the deployment may have been required to be from a FaradayRF repository branch. The following output were the last two lines in the build:

Skipping a deployment with the pypi provider because this branch is not permitted
Done. Your build exited with 0.

This appears to have at least triggered an automated deployment of an otherwise good module.