There are no "next" arrow or buttons on any page. The user has to go back to the main page every time to select the next chapter.

Great to see a workshop in progress on Python testing and CI! I'm curious whether including a short section on how to use tools like pylint and black might fit in well here. They're be out of scope if you really want to keep the focus on testing, but they'd fit in with the broader bucket of useful tools for maintaining quality code that can be used with CI.

Is there value in adding material to cover simple test parameterisation?

I imagine this could fit in after edge and corner cases, reuse the sinc2d example, and make use of the @pytest.mark.parametrize decorator (docs are at https://docs.pytest.org/en/latest/parametrize.html). Essentially this could be "something else useful that pytest does" and could be a task of refactoring the collection of edge and corner case tests.

As I see it there are two down sides of this. First, it would make the lesson longer, and that's not necessarily a good thing. Second, it would make use of decorators, and these would probably be new to many learners.

In 04-units, it would be good to add a test for an empty list - unless it is too advanced to test for catching exceptions?

This lesson used to be listed at http://software-carpentry.org/lessons/ (I think) but is no longer linked from there. It would be a good candidate for the Carpentries Incubator in opinion. Please consider adding it there, see https://github.com/carpentries-incubator/proposals.

Hi Katy

I have been in contact with Greg and Jonah from Software Carpentry, as I thought to present a Unit Testing course to the scientists in my region (and possibly punt your book!). They said that the course you have prepared is great, although suggested that I could contribute by bringing the template and styling up to date. Would you be ok with that ? Greg said he would be able to help me along with getting through what needs to be altered and that it shouldn't be a huge amount of work.

Sincerely
Liam

In http://katyhuff.github.io/python-testing/05-pytest/, there is an exercise at the end to make all tests pass by ensuring mean.py handles complex number by returning NotImplemented. What is the preferred solution? I came up with adding this at the beginning:

    if isinstance(sum(num_list), complex):
        return NotImplemented

This works, but causes test_long to take much more time (due to the summing of 100000000 numbers).

(@annefou and I'll be teaching this workshop in two weeks, hence the question)

The latest version of pytest has pytest as its command. py.test still works, but it may still be worth to update the lesson on this point.

I taught this lesson yesterday as part of a more-advanced SWC workshop for 30. It's great! While I remember one common problem the students hit in the Continuous Integration part is that travisCI is very picky about the format of the .travis.yaml file. It's probably worth adding a note about this either in the lesson or in the instructor notes.

A useful tool to point to would be the travis.yaml linter at https://lint.travis-ci.org/ and maybe note some common error cases somewhere. These seem to be:

• .travis.yaml is malformed. There is an error message at the top of the job log, and pointing to the linter is probably the best plan.
• the tests fail and end with a message about "rake". This is an attempt to run tests for Ruby (the default case for Travis) and this probably means that the .travis.yaml has not been found. A common reason seems to be that people forget the first dot at the start of the filename. The fix is to rename (git mv ..., git commit, git push) and try again.
• the tests don't run. This happens if the student adds the files and commits them before turning on tests for the repository in travis (I think there is a bit of a delay between linking the travis and github account and being able to do anything, so if the student moves on to keep up they may do the push before flipping the switch in travis). The fix is to push some other change to the repo.

Anyway, I'll try to turn these into a pull request when I get a chance, but at least if I stick this here I'll not completely forget.

Lesson 5 "Running Tests with pytest" is absent from this repo.