Code coverage about python-tuf HOT 5 CLOSED

I believe this is the goal of the system tests Konstantin is currently working on. It appears they are proving useful, as they are indeed hitting code paths that would have been trickier to catch otherwise. The new tuf-pypi mirror will also help tremendously in this aspect. Hopefully we could put in some solid testing now.

I think our tests have been thorough enough so far. They are still in the process of being implemented, but once they are complete we can go back and address any issues that may still exist.

from python-tuf.

Agreed. I also meant that we should use tools like Ned Batchelder's coverage, which will give us a precise way of measuring exactly how much code is "covered".

I learned about things like these from the Rails people: they have some good practices, such as code coverage, integration tests and test-driven development.

from python-tuf.

While we are it, we should also look at nose.

from python-tuf.

Finally, at some point, it would be sweet to automatically run nose tests (with the coverage.py plugin) with Travis every time we commit to GitHub.

The pip folks do this continuous integration.

from python-tuf.

The Update Framework now uses Coverage (including branch coverage), Travis, and Tox.

from python-tuf.