This test was written using an incorrect assumption that the
pillar_roots key created during git_pillar execution lives any longer
than the run that compiles the git_pillar data for a given git_pillar.
This test was written using an incorrect assumption that the
pillar_roots key created during git_pillar execution lives any longer
than the run that compiles the git_pillar data for a given git_pillar.
If module named git does not exist, test test_ext_pillar_env_mapping
fail:
======================================================================
ERROR: test_ext_pillar_env_mapping
(integration.modules.pillar.PillarModuleTest)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/opt/git/salt/tests/integration/modules/pillar.py", line 120, in
test_ext_pillar_env_mapping
import git
ImportError: No module named git
Added skip test_ext_pillar_env_mapping test.
Tests in pillar.py are importing the "git" module, which is supposed to
be "GitPython", bt since Python 2.x doesn't use absolute import by
default, it imports the "git" submodule, which defines the tests for the
Git module of Salt, instead.
Defined "integration.run_tests()" function which can be used to execute
a particular integration test case. Existing bolerplate code in modules
and states tests is replaced with following lines:
if __name__ == '__main__':
from integration import run_tests
run_tests(TestCaseName)
Typical usecase could look like this:
python integration/modules/pip.py --no-clean -vv
