Reaping test: remove no sleep assumption

The reaping test assumes that all the `sleep` processes on the system
are created by us, which is an incorrect assumption, and creates
problems now that we run in a more "complete" system in CI (i.e. not in
a minimal container).

This solves the issue by not assuming that random sleep processes are
related to us: instead, we look at whether these processes are in our
process group.

This also includes a little clean up of the reaping stage 1.

test/reaping/stage_1.py

@@ -8,33 +8,50 @@ import time
 import psutil
 
 
+def in_group_or_reaped(pid):
+    try:
+        return os.getpgid(pid) == os.getpgid(0)
+    except OSError:
+        return True
+
+
 def main():
-    p = subprocess.Popen([os.path.join(os.path.dirname(__file__), "stage_2.py")])
-    p.wait()
+    stage_2 = os.path.join(os.path.dirname(__file__), "stage_2.py")
+    subprocess.Popen([stage_2]).wait()
 
     # In tests, we assume this process is the direct child of init
     this_process = psutil.Process(os.getpid())
     init_process = this_process.parent()
 
-    print("Reaping test: stage_1 is pid{0}, init is pid{1}".format(this_process.pid, init_process.pid))
+    print("Reaping test: stage_1 is pid{0}, init is pid{1}".format(
+        this_process.pid, init_process.pid))
 
     # The only child PID that should persist is this one.
     expected_pids = [this_process.pid]
 
-    print("Expecting pids to remain: {0}".format(", ".join(str(pid) for pid in expected_pids)))
+    print("Expecting pids to remain: {0}".format(
+        ", ".join(str(pid) for pid in expected_pids)))
 
     while 1:
         pids = [p.pid for p in init_process.children(recursive=True)]
         print("Has pids: {0}".format(", ".join(str(pid) for pid in pids)))
+        for pid in pids:
+            assert in_group_or_reaped(pid), "Child had unexpected pgid"
         if set(pids) == set(expected_pids):
             break
         time.sleep(1)
 
-    # Now, check if there are any zombies
+    # Now, check if there are any zombies. For each of the potential zombies,
+    # we check that the pgid is ours.  NOTE: We explicitly test that this test
+    # fails if subreaping is disabled, so we can be confident this doesn't turn
+    # a failure into a success.
     for process in psutil.process_iter():
-        if process.name() == "sleep":
-            print("At least one 'sleep' process was still alive or not reaped! (pid{0})".format(process.pid))
-            sys.exit(1)
+        if process.pid == this_process.pid:
+            continue
+        if not in_group_or_reaped(process.pid):
+            continue
+        print("Not reaped: pid {0}: {1}".format(process.pid, process.name()))
+        sys.exit(1)
 
     sys.exit(0)