I want to try running the unit tests on Windows (native), but first, I thought it wise to run them on Ubuntu to make sure I knew how to run them, and to get a baseline idea of how many tests are supposed to run and whether they are all passing. The first problem I'm hitting is that the suite stalls after 302 tests at test_extract_examples, railing out a CPU core and running indefinitely (at least a couple minutes before I killed it).
Before I attempt to debug the test---which employs some amount of indirection to call out to a main function---I wanted to ask about two things:
1. Is this a known or expected problem?
2. Under what conditions/environment do I need to run the test? Or, Martin, what do you do, and how flexible is that?
In a nutshell (details later), my current setup involves a proper Ubuntu 17.10 in a VM (not WSL), installing from source into a virtualenv, and then invoking nosetests against the installation (not the repo). Trying it against the repo yielded a bunch of errors that I did not investigate for now. Is this a reasonable arrangement? What is the appropriate setup for running these tests?
Details:
Ubuntu 17.10 Desktop 64-bit running in VMWare Player
Existing packages updated.
Installed Python: 3.6.3
sudo apt install python3-venv python3-pip python3-dev
Setup a python virtualenv using python3 -m venv ENVNAME
With the virtual environment activated:
pip install wheel (to avoid gripes about unable to build local wheels of dependencies)
pip install nose
cd path/to/repo/beancount
pip install .
cd path/to/elsewere
nosetests -v beancount