Automated Pint Testing

[hgrecco]

I have created a repository to make sure we do not break backwards compatibility unintentionally in Pint.

Take a look at https://github.com/hgrecco/pint_downstream_tester

[rkingsbury]

Awesome; thanks for setting this up! Is it working? I took a peek at your travis.yml file, and on my system at least, I have to change

- python -m unittest pyEQL/tests/*

to

- python -m unittest pyEQL/tests/*.py

To get the tests to run.

[hgrecco]

It seems to work the way it is:
https://travis-ci.org/hgrecco/pint_downstream_tester/builds/116230091

But in any case, if I may suggest it, I would be nice to have a better entry point for testing. For example, in Pint we do the following:

In the test suite we have function to collect tests and run them. See here

This functions is called by a function named test in the package __init__.py See here

In this way users can just run:

>>> from pint import test
>>> test()

The final piece of the puzzle is a line in setup.py. See here
which allows users to run in the console

python setup.py test

Of course, not all this is necessary. But is nice to have at least a part of it to make it easier for end users to run tests (and for crazy people like me having this kind of integrated testing!).

[rkingsbury]

As a matter of fact a better entry point for testing has been on my "to-do" list, I just haven't had time to research a better way to set it up. Looks like you've given me the solution I was looking for; thank you! I'll implement this before the next release.