Skip to main content

testrig: running tests for dependent packages

Project description


Python package integration testing.

Runs the test suite of some package against ‘old’ and ‘new’ versions of given dependencies. Failures that appear in ‘new’ are reported.

Each test suite run is run in a virtualenv constructed from scratch. You should install ccache (and possibly also f90cache) to avoid drinking too much coffee.

Alternatively, binary conda packages can be used — however, binary incompatibities may arise in this configuration.

Currently, this is POSIX-only, and tested only on Linux.

See Travis-CI for most recent results for latest Numpy maintenance branch. Log in to Travis and press the restart button to re-run the checks for the latest commit / tag.



python --help
python pandas                            # run tests, default config
python --config=testrig-conda.ini pandas # use conda packages
python -j                                # run all packages parallel

The runs may take a long time, as it builds everything from source.


Configuration is read from testrig.ini by default. It contains sections, one per test environment. Section named DEFAULT can be used to specify (overridable) default values for the configuration items.

An example first (runs scipy test suite against old and new numpy versions):

new=Cython==0.22 git+

base=nose tempita Cython==0.22 scipy==0.17.0
run=python -c 'import numpy; numpy.test("fast", verbose=2)'

The configuration items in each section are:

  • env: which environment to use
    • virtualenv: virtualenv + pip, all packages are built from sources
    • conda: conda, uses binary packages, except for git+ urls and package names prefixed by pip+. Note that you may need to write stuff like numpy git+ since conda only understand that packages installed by it are present.
  • old: package specifications for the ‘old’ configuration.
  • new: package specifications for the ‘new’ configuration.
  • base: packages for both configurations. These are installed after those specified by old or new.
  • run: command that runs the tests.
  • parser: parser for the test output. Available options:
    • junit: parses xUnit/jUnit xml result output. It is assumed the result is stored in a file junit.xml in the working directory. Both nose and py.test can produce it: py.test --junit-xml=junit.xml ... and nosetests --with-xunit --xunit-file=junit.xml ....
    • nose: parses nose stdout
    • pytest-log: parses contents from py.test --result-log=pytest.log .... The results are assumed to be stored in pytest.log in the current working directory.
  • envvars: additional environment variables to set (also for pip install). The text $DIR is replaced by an absolute path of the directory where the configuration file resides.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

testrig-0.1.1.tar.gz (92.8 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page