Reasonable wercker config

[grzanka]

Prepare more reasonable wercker config and add finally badge to README.rst
To be discussed here.

[antnieszka]

There are 2 available badges:

and

There is also this setting: Show the status of all branches or only "master" with all and master options.

[grzanka]

smaller is better, I'd also prefer only master

[antnieszka]

Ok. Should we use one Python version in wercker box or try to run tests for multiple? I'm investigating how we could the latter. There are some options:

• https://github.com/flenter/multi-python-example/blob/master/wercker.yml
• pyenv https://github.com/nabetama/box-pyenv/blob/master/wercker-box.yml

And... maybe the most important question - what kind of tests, or what at all could we run on wercker? It's really quick - comparing to travis.

[antnieszka]

I think we could create our own docker file and use it with wercker.

[grzanka]

First of all I would like to use workflows. If some stage fails, then it doesn't make to move further.
For example:

1. Try to make a package, if this fails, doesn't make sense to go further
2. Check if code is correct PEP8 formatted, if yes move further
3. Check if tests pass on Python3.4 and Python2.7 (in parallel if possible), if yes, move further
4. Check other versions of python (in parallel if possible): 3.2,3.3,3.4,3.6
5. Run other tests
6. Run some integration tests (i.e. create new docker instance (or many using different Linux distros), install beprof using "pip+git" method, and run some example code which makes use of the library (issue Profile examples using urllib and numpy file reading #9))

Can it be done in wercker ?

[antnieszka]

I'll try to make something like Proof of Concept based on steps above.

[antnieszka]

If you don't mind - I'll test it on my repo fork, because it will be probably a lot of trash.

[grzanka]

Sure

[antnieszka]

@grzanka I have another minor decision to make:

Processing ./beprof-0.post0.dev2111480854429-py2.py3-none-any.whl
Collecting numpy>=1.10.4 (from beprof==0.post0.dev2111480854429)
  Downloading numpy-1.11.2-cp27-cp27mu-manylinux1_x86_64.whl (15.3MB)
Installing collected packages: numpy, beprof
  Found existing installation: numpy 1.8.0
    DEPRECATION: Uninstalling a distutils installed project (numpy) has been deprecated and will be removed in a future version. This is due to the fact that uninstalling a distutils project will only partially uninstall the project.
    Uninstalling numpy-1.8.0:
      Successfully uninstalled numpy-1.8.0
Successfully installed beprof-0.post0.dev2111480854429 numpy-1.11.2

It looks like repo-installed numpy is being replaced when installing beprof with pip. So the whole "Testing repo installed numpy" is pointless in this case. I'll for now use pip install --no-deps beprof*.whl.

Edit: more can be found here https://app.wercker.com/ant6/beprof/runs/test_opensuse_leap_repo/58440ba485198d010082fa9e?step=58440bb3c6f72d000173cd79

[antnieszka]

Checklist here: #105