Fix DeprecationWarning about fork() and multi-threaded processes in…

… unit tests

Ensure all threads are properly shutting down in the unit tests. This prevents the following warning:
```
DeprecationWarning: This process (pid=208847) is multi-threaded, use of fork() may lead to deadlocks in the child.
```

conftest.py

@@ -0,0 +1,10 @@
+import threading
+
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def assert_no_threads_left():
+    assert len(threading.enumerate()) == 1
+    yield
+    assert len(threading.enumerate()) == 1

libmuscle/python/libmuscle/manager/test/conftest.py

@@ -36,7 +36,7 @@ def mmp_configuration():
 def profile_store(tmp_path):
     test_profile_store = ProfileStore(tmp_path)
     yield test_profile_store
-    test_profile_store.close()
+    test_profile_store.shutdown()
 
 
 @pytest.fixture

libmuscle/python/libmuscle/manager/test/test_profile_store.py

@@ -9,7 +9,7 @@
 
 def test_create_profile_store(tmp_path):
     db = ProfileStore(tmp_path)
-    db.close()
+    db.shutdown()
 
     db_path = tmp_path / 'performance.sqlite'
     conn = sqlite3.connect(db_path, isolation_level=None)
@@ -114,4 +114,4 @@ def check_register_event(typ, start, stop):
 
     cur.close()
     conn.close()
-    db.close()
+    db.shutdown()

libmuscle/python/libmuscle/test/test_instance.py

@@ -142,23 +142,26 @@ def instance_dont_apply_overlay(
 def test_create_instance_manager_location_default(
         instance_argv, MMPClient, declared_ports):
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
+    instance.error_shutdown("")  # ensure all threads and resources are cleaned up
 
     MMPClient.assert_called_once_with(Ref('component'), 'tcp:localhost:9000')
 
 
 def test_create_instance_manager_location_argv(
         manager_location_argv, instance_argv, MMPClient, declared_ports):
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
+    instance.error_shutdown("")  # ensure all threads and resources are cleaned up
 
     MMPClient.assert_called_once_with(Ref('component'), 'tcp:localhost:9001')
 
 
 def test_create_instance_manager_location_envvar(
         manager_location_envvar, instance_envvar, MMPClient, declared_ports):
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
+    instance.error_shutdown("")  # ensure all threads and resources are cleaned up
 
     MMPClient.assert_called_once_with(Ref('component[13]'), 'tcp:localhost:9002')
 
@@ -167,7 +170,8 @@ def test_create_instance_registration(
         manager_location_argv, instance_argv, mmp_client, communicator, port_manager,
         profiler, declared_ports):
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
+    instance.error_shutdown("")  # ensure all threads and resources are cleaned up
 
     locations = communicator.get_locations.return_value
 
@@ -195,9 +199,10 @@ def test_create_instance_registration(
 def test_create_instance_profiling(
         manager_location_argv, instance_argv, profiler, declared_ports):
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
 
     assert len(profiler.record_event.mock_calls) == 2
+    instance.error_shutdown("Ensure all threads and resources are cleaned up")
 
 
 def test_create_instance_connecting(
@@ -214,7 +219,7 @@ def test_create_instance_connecting(
     settings = MagicMock()
     mmp_client.get_settings.return_value = settings
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
 
     port_manager.connect_ports.assert_called_once()
     peer_info = port_manager.connect_ports.call_args[0][0]
@@ -227,13 +232,14 @@ def test_create_instance_connecting(
     assert communicator.set_peer_info.call_args[0][0] == peer_info
 
     assert settings_manager.base == settings
+    instance.error_shutdown("Ensure all threads and resources are cleaned up")
 
 
 def test_create_instance_set_up_checkpointing(
         manager_location_argv, instance_argv, mmp_client, trigger_manager,
         no_resume_snapshot_manager, settings_manager, declared_ports):
 
-    Instance(declared_ports)
+    instance = Instance(declared_ports)
 
     elapsed_time, checkpoints, resume_path, snapshot_path = (
             mmp_client.get_checkpoint_info.return_value)
@@ -242,6 +248,7 @@ def test_create_instance_set_up_checkpointing(
     no_resume_snapshot_manager.prepare_resume.assert_called_with(
             resume_path, snapshot_path)
     assert settings_manager.overlay != no_resume_snapshot_manager.resume_overlay
+    instance.error_shutdown("Ensure all threads and resources are cleaned up")
 
 
 def test_create_instance_set_up_logging(
@@ -272,6 +279,7 @@ def get_logger(name=''):
         root_logger.setLevel.assert_called_with(logging.ERROR)
         libmuscle_logger.setLevel.assert_called_with(logging.DEBUG)
         ymmsl_logger.setLevel.assert_called_with(logging.DEBUG)
+    instance.error_shutdown("Ensure all threads and resources are cleaned up")
 
 
 def test_shutdown_instance(
@@ -540,4 +548,5 @@ def test_checkpoint_support(
     mmp_client.get_checkpoint_info.return_value = checkpoint_info
 
     with expectation:
-        Instance(flags=flags)
+        instance = Instance(flags=flags)
+        instance.error_shutdown("Ensure all threads and resources are cleaned up")