Paralelize the integration tests

This is an attempt to fix fedora-python#45

It does the following:

 * uses pytest-xdist to paralelize the tests (3 workers)
 * utilizes custom scheduler that splits the tests acroding to the fixture name
   * this is needed not to use 1 fixture on multiple workers
     see pytest-dev/pytest-xdist#18
 * use parametrize an all integration tests to enbale our hackish scheduler
 * mock now creates the roots in pwd (not to pollute the filesystem on /)
   * note that the roots can take sveral GBs

Author: hroncok

.gitignore

@@ -6,6 +6,7 @@ __pycache__
 /artifacts-*/
 /build/
 /dist/
+/mockroots/
 /.tox/
 /.eggs/
 .cache

mock.cfg

@@ -4,4 +4,5 @@ config_opts['chroot_setup_cmd'] = 'install ansible dnf'
 config_opts['use_host_resolv'] = True
 config_opts['rpmbuild_networking'] = True
 config_opts['use_nspawn'] = False
-config_opts['root'] = 'fedora-27-x86_64-taskotron'
+config_opts['root'] = 'taskotron-python-versions-master'
+config_opts['plugin_conf']['root_cache_opts']['dir'] = "%(cache_topdir)s/taskotron-python-versions/root_cache/"

test/integration/conftest.py

@@ -1,4 +1,22 @@
+from xdist.scheduler import LoadScopeScheduling
+
+
 def pytest_addoption(parser):
     parser.addoption('--fake', action='store_true', default=False,
                      help='don\'t run the code, reuse the result from '
                           'last tests')
+
+
+class FixtureScheduling(LoadScopeScheduling):
+    """Split by [] value. This is very hackish and might blow up any time!
+    See https://github.com/pytest-dev/pytest-xdist/issues/18
+    """
+    def _split_scope(self, nodeid):
+        if '[' in nodeid:
+            parameters = nodeid.rsplit('[')[-1].replace(']', '')
+            return parameters.split('-')[0]
+        return None
+
+
+def pytest_xdist_make_scheduler(log, config):
+    return FixtureScheduling(config, log)

test/integration/test_integration.py

@@ -1,6 +1,7 @@
 from collections import namedtuple
 import contextlib
 import glob
+import os
 import pprint
 import shutil
 import subprocess
@@ -16,14 +17,27 @@
 
 
 class MockEnv:
-    '''Use this to work with mock. Mutliple instances are not safe.'''
+    '''Use this to work with mock. Mutliple concurrent instances are safe.'''
     mock = ['mock', '-r', './mock.cfg']
 
-    def __init__(self):
+    def __init__(self, worker_id):
+        self.worker_id = worker_id
         self._run(['--init'], check=True)
 
+    @property
+    def root(self):
+        return 'taskotron-python-versions-{}'.format(self.worker_id)
+
+    @property
+    def rootdir(self):
+        return os.path.join(os.path.abspath('.'), 'mockroots', self.root)
+
     def _run(self, what, **kwargs):
-        return subprocess.run(self.mock + what, **kwargs)
+        command = list(self.mock)  # needs a copy not to change in place
+        command.append('--config-opts=root={}'.format(self.root))
+        command.append('--rootdir={}'.format(self.rootdir))
+        command.extend(what)
+        return subprocess.run(command, **kwargs)
 
     def copy_in(self, files):
         self._run(['--copyin'] + files + ['/'], check=True)
@@ -52,12 +66,12 @@ def copy_out(self, directory, target, *, clean_target=False):
 
 
 @pytest.fixture(scope="session")
-def mock(request):
+def mock(worker_id, request):
     '''Setup a mock we can run Ansible tasks in under root'''
     if request.config.getoption('--fake'):
-        mockenv = FakeMockEnv()
+        mockenv = FakeMockEnv(worker_id)
     else:
-        mockenv = MockEnv()
+        mockenv = MockEnv(worker_id)
     files = ['taskotron_python_versions'] + glob.glob('*.py') + ['tests.yml']
     mockenv.copy_in(files)
     yield mockenv
@@ -187,8 +201,10 @@ def test_two_three_passed(results, request):
     assert results['dist.python-versions.two_three'].outcome == 'PASSED'
 
 
-def test_two_three_failed(tracer):
-    assert tracer['dist.python-versions.two_three'].outcome == 'FAILED'
+@pytest.mark.parametrize('results', ('tracer',))
+def test_two_three_failed(results, request):
+    results = request.getfixturevalue(results)
+    assert results['dist.python-versions.two_three'].outcome == 'FAILED'
 
 
 @pytest.mark.parametrize('results', ('tracer', 'copr', 'admesh'))
@@ -207,8 +223,10 @@ def test_artifact_is_the_same(results, task, request):
             results['dist.python-versions.' + task].artifact)
 
 
-def test_artifact_contains_two_three_and_looks_as_expected(tracer):
-    result = tracer['dist.python-versions.two_three']
+@pytest.mark.parametrize('results', ('tracer',))
+def test_artifact_contains_two_three_and_looks_as_expected(results, request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.two_three']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -233,8 +251,11 @@ def test_naming_scheme_failed(results, request):
     assert results['dist.python-versions.naming_scheme'].outcome == 'FAILED'
 
 
-def test_artifact_contains_naming_scheme_and_looks_as_expected(copr):
-    result = copr['dist.python-versions.naming_scheme']
+@pytest.mark.parametrize('results', ('copr',))
+def test_artifact_contains_naming_scheme_and_looks_as_expected(results,
+                                                               request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.naming_scheme']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -258,9 +279,11 @@ def test_requires_naming_scheme_failed(results, request):
     assert task_result.outcome == 'FAILED'
 
 
+@pytest.mark.parametrize('results', ('tracer',))
 def test_artifact_contains_requires_naming_scheme_and_looks_as_expected(
-        tracer):
-    result = tracer['dist.python-versions.requires_naming_scheme']
+        results, request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.requires_naming_scheme']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -281,8 +304,10 @@ def test_artifact_contains_requires_naming_scheme_and_looks_as_expected(
     """).strip() in artifact.strip()
 
 
-def test_requires_naming_scheme_contains_python(yum):
-    result = yum['dist.python-versions.requires_naming_scheme']
+@pytest.mark.parametrize('results', ('yum',))
+def test_requires_naming_scheme_contains_python(results, request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.requires_naming_scheme']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -306,9 +331,11 @@ def test_executables_failed(results, request):
     assert task_result.outcome == 'FAILED'
 
 
+@pytest.mark.parametrize('results', ('docutils',))
 def test_artifact_contains_executables_and_looks_as_expected(
-        docutils):
-    result = docutils['dist.python-versions.executables']
+        results, request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.executables']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -352,9 +379,11 @@ def test_unvesioned_shebangs_failed(results, request):
     assert result.outcome == 'FAILED'
 
 
+@pytest.mark.parametrize('results', ('tracer',))
 def test_artifact_contains_unversioned_shebangs_and_looks_as_expected(
-        tracer):
-    result = tracer['dist.python-versions.unversioned_shebangs']
+        results, request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.unversioned_shebangs']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -378,9 +407,11 @@ def test_unvesioned_shebangs_mangled_failed(results, request):
     assert result.outcome == 'FAILED'
 
 
+@pytest.mark.parametrize('results', ('bucky',))
 def test_artifact_contains_mangled_unversioned_shebangs_and_looks_as_expected(
-        bucky):
-    result = bucky['dist.python-versions.unversioned_shebangs']
+        results, request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.unversioned_shebangs']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -420,15 +451,17 @@ def test_py3_support_failed(results, request):
     assert task_result.outcome == 'FAILED'
 
 
+@pytest.mark.parametrize('results', ('bucky',))
 def test_artifact_contains_py3_support_and_looks_as_expected(
-        bucky):
+        results, request):
     """Test that py3_support check fails if the package is mispackaged.
 
     NOTE: The test will start to fail as soon as python-bucky
     gets ported to Python 3 and its Bugzilla gets closed.
     See https://bugzilla.redhat.com/show_bug.cgi?id=1367012
     """
-    result = bucky['dist.python-versions.py3_support']
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.py3_support']
     with open(result.artifact) as f:
         artifact = f.read()
 
@@ -459,8 +492,11 @@ def test_python_usage_failed(results, request):
     assert task_result.outcome == 'FAILED'
 
 
-def test_artifact_contains_python_usage_and_looks_as_expected(jsonrpc):
-    result = jsonrpc['dist.python-versions.python_usage']
+@pytest.mark.parametrize('results', ('jsonrpc',))
+def test_artifact_contains_python_usage_and_looks_as_expected(results,
+                                                              request):
+    results = request.getfixturevalue(results)
+    result = results['dist.python-versions.python_usage']
     with open(result.artifact) as f:
         artifact = f.read()
 

tox.ini

@@ -14,13 +14,14 @@ sitepackages = True
 [testenv:integration]
 deps =
     pytest
+    pytest-xdist
     pyyaml
 basepython = python3
-commands = python -m pytest -v {posargs} test/integration
+commands = python -m pytest -v -n3 {posargs} test/integration
 sitepackages = False
 
 [testenv:style]
 deps = flake8
 basepython = python3
-commands = python -m flake8 . --ignore=E402
+commands = python -m flake8 . --ignore=E402 --exclude=.git,__pycache__,.tox,.eggs,dist,build,mockroots
 sitepackages = False