Add tests for tracing down the issue with bluetoothctl version

As it turned out, there was a regression in Python 3.11.1
related to loss of output from subprocesses.

Refs: hbldh/bleak#1183
      python/cpython#100133

tests/run_bleak_version_mp.py

@@ -0,0 +1,28 @@
+import asyncio
+import multiprocessing
+import os
+
+import bleak.backends.bluezdbus.version as bv
+
+async def subprocess_task():
+    version_output  = await bv._get_bluetoothctl_version()
+    if version_output:
+        major, minor = tuple(map(int, version_output.groups()))
+        print(major, minor)
+    else:
+        print("No version_output")
+    await asyncio.sleep(1.0)
+
+def subprocess_work():
+    sploop = asyncio.new_event_loop()
+    print(f"Inner loop: {sploop} {os.getpid()} {sploop._selector}")
+    sploop.run_until_complete(subprocess_task())
+
+
+if __name__ == '__main__':
+    multiprocessing.set_start_method('spawn', force=True)
+    loop = asyncio.new_event_loop()
+    print(f"Outer loop: {loop} {os.getpid()} {loop._selector}")
+    loop.run_until_complete(subprocess_task())
+    p = multiprocessing.Process(target=subprocess_work)
+    p.start()

tests/test_bleak_version.py

@@ -0,0 +1,14 @@
+import bleak.backends.bluezdbus.version as bv
+
+import pytest
+
+@pytest.mark.asyncio
+async def test_version():
+    version_output  = await bv._get_bluetoothctl_version()
+    assert version_output
+    major, minor = tuple(map(int, version_output.groups()))
+    assert major == 5
+    assert minor == 55
+
+
+