Opened 10 years ago
Closed 9 years ago
#234 closed defect (fixed)
test failure on Fedora in test_tub.CancelPendingDeliveries.test_cancel_pending_deliveries
Reported by: | davidsarah | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | 0.9.0 |
Component: | unknown | Version: | 0.6.4 |
Keywords: | test | Cc: | davidsarah |
Description
See https://bugzilla.redhat.com/show_bug.cgi?id=1106775. Almost all of the test failures there are due to #141, except for this one:
[ERROR] Traceback (most recent call last): File "/builddir/build/BUILD/foolscap-0.6.4/foolscap/test/test_tub.py", line 386, in tearDown if self.tubA.running: exceptions.AttributeError: 'CancelPendingDeliveries' object has no attribute 'tubA' foolscap.test.test_tub.CancelPendingDeliveries.test_cancel_pending_deliveries ==================================
Change History (3)
comment:1 Changed 10 years ago by
Cc: | davidsarah added |
---|---|
Version: | 0.7.0 → 0.6.4 |
comment:2 Changed 10 years ago by
comment:3 Changed 9 years ago by
Resolution: | → fixed |
---|---|
Status: | new → closed |
I haven't heard anything more about this, so I'm going to assume that 0.8.0 fixed it as expected.
Note: See
TracTickets for help on using
tickets.
Oh, if SSL was failing on that test run, this this test would fail as an artifact. That exception means that
teardown()
was called before the test finished its first line (which createsself.tubA
. If the constructor forGoodEnoughTub()
failed due to SSL problems, then the test would error out (without settingself.tubA
), and jump immediately toteardown()
, which would then fail to handle the attribute that never got created.So I think this is a consequence of #141 too. I believe it will be gone with the SSL fix that went into 0.8.0.