[Zope-Checkins] CVS: ZODB/src/ZODB/tests - PackableStorage.py:1.34

[Tim Peters]

Update of /cvs-repository/ZODB/src/ZODB/tests
In directory cvs.zope.org:/tmp/cvs-serv11782/src/ZODB/tests

Modified Files:
	PackableStorage.py 
Log Message:
_PackWhileWriting:  Cut the number of threads and boost the max timeout.
This makes the ZEO flavors of the tests much more likely to complete
without a "thread still alive" failure.


=== ZODB/src/ZODB/tests/PackableStorage.py 1.33 => 1.34 ===
--- ZODB/src/ZODB/tests/PackableStorage.py:1.33	Fri Mar 12 13:08:13 2004
+++ ZODB/src/ZODB/tests/PackableStorage.py	Fri Mar 12 14:41:50 2004
@@ -175,10 +175,20 @@
                 root[i].value = MinPO(i)
                 get_transaction().commit()
 
+        # How many client threads should we run, and how long should we
+        # wait for them to finish?  Hard to say.  Running 4 threads and
+        # waiting 30 seconds too often left a thread still alive on Tim's
+        # Win98SE box, during ZEO flavors of this test.  Those tend to
+        # run one thread at a time to completion, and take about 10 seconds
+        # per thread.  There doesn't appear to be a compelling reason to
+        # run that many threads.  Running 3 threads and waiting up to a
+        # minute seems to work well in practice.  The ZEO tests normally
+        # finish faster than that, and the non-ZEO tests very much faster
+        # than that.
         NUM_LOOP_TRIP = 50
         timer = ElapsedTimer(time.time())
         threads = [ClientThread(db, choices, NUM_LOOP_TRIP, timer, i)
-                   for i in range(4)]
+                   for i in range(3)]
         for t in threads:
             t.start()
 
@@ -188,7 +198,7 @@
             db.pack(packt)
 
         for t in threads:
-            t.join(30)
+            t.join(60)
         liveness = [t.isAlive() for t in threads]
         if True in liveness:
             # They should have finished by now.