Actually, seewind's code uses FileStorage, too. I modified seewind's test_db.py to leave out the use of the process module, which I don't have, and I was able to reproduce the errors. It isn't clear to me exactly what is distinct about this test that triggers the "KeyError: 'pop from an empty set'" exception, but it appears to be corrected by changing the line in get_packer() from while len(self.pack_extra) > n: first.append(self.pack_extra.pop()) to while self.pack_extra and len(self.pack_extra) > n: first.append(self.pack_extra.pop()) I'll check in that change. Thanks, seewind, for reporting this. The other type of error seewind reports is a ConflictError. A ConflictError may not indicate a bug, but we need to figure out just what is going on in this case to be sure. On Mar 8, 2009, at 2:18 PM, Michael Watkins wrote: > On Thu, February 19, 2009 11:46 pm, seewind wrote: >> the test code can be download at: >> http://seewind.googlecode.com/files/durus_test.zip >> I am run run.bat at windows, it will raise some errors. errors >> message >> note at errormessage.txt. I found do not run "pack()", will be ok. > > As you are using a third party storage - the bsddb storage - it would > probably be best if someone familiar with that package jumped in here. > > > _______________________________________________ > Durus-users mailing list > Durus-users@mems-exchange.org > http://mail.mems-exchange.org/mailman/listinfo/durus-users