At least one of the tests requires the existence of the testbed. But we'd like to be able to run as many tests as possible from within the conda build itself. To that end I'm marking testbed-dependent tests with pytest.mark.testbed.
The two tests are:
test_config.py/test_configuration: this test is designed to verify that the testbed contains a proper mix of kernel installation scenarios.
test_config.py/test_install_kernelspec: this test will fail unless there is at least one kernel installed outside of the conda-build environment.
Because the CI does exercise these tests fully, we are safe to skip them in the recipe build.
In order to test this, we've removed the --no-test option in the build stage.
At least one of the tests requires the existence of the testbed. But we'd like to be able to run as many tests as possible from within the conda build itself. To that end I'm marking testbed-dependent tests with
pytest.mark.testbed
.The two tests are:
test_config.py/test_configuration
: this test is designed to verify that the testbed contains a proper mix of kernel installation scenarios.test_config.py/test_install_kernelspec
: this test will fail unless there is at least one kernel installed outside of the conda-build environment.Because the CI does exercise these tests fully, we are safe to skip them in the recipe build.
In order to test this, we've removed the
--no-test
option in thebuild
stage.