Avi Miller wrote:
% Hi,
%
% Please ignore this patch. There are other changes required to support the
Oracle Linux 6 key, which I'm still researching. Apologies for the noise.
%
% Thanks,
% Avi
Hello Avi,
as we discussed on IRC - currently spacewalk doesn't correctly recognize
signature type
Avi Miller wrote:
% There is an issue with the default yum repo label with some of the new
repositories, particularly the Scientific Linux and Oracle Linux Spacewalk
Server/Client repos. The full text expands to greater than 64 chars, which is
the limit of the database field that stores the
Hi,
We propose some patches to avoid the junit testsuite to depend on
Cobbler. With those, we are able to run all the tests using only a
database instance seeded with the latest schema version. Details:
- patch 17 calls setupTestConfiguration() on KickstartDataTest before
calling other members
Hi,
As previously discussed the new logging feature requires calling
LoggingFactory.clearLogId() at the beginning of each transaction, and
there were some places in test code where that was missing.
I confirm that the latest patches from Grant actually covered most
cases, I could only find four
Hi,
In some frontend tests it is assumed that at least one Monitoring Scout
exists, and those tests will fail if there is none.
The proposed attached patch creates a test Monitoring Scout before any
operation that requires one, thus eliminating this precondition.
Regards,
--
Silvio Moioli
SUSE
Hi,
If I understand code correctly MasterHandlerTest is about testing that
some methods can or cannot be called by administrators and regular
users. IssMaster objects are created ad-hoc for testing, and one of them
is flagged and unflagged as default during the tests themselves. The
only problem
Hi,
The attached patch fixes a previous contribution of mine to
transparently get resources either from the filesystem or from a jar
file. When accessing a file from a jar that files gets decompressed in a
temporary location which name was generated at random for each call.
This however did not
SystemHandlerTest was failing in our environment and, AFAIU, it expects
that the underlying database will always return 1 as the id of a
persisted test object. This is normally not the case, as the value is
taken from a sequence and there is no code to ensure it is reset before
the test, so I
See attached patch.
Regards,
--
Silvio Moioli
SUSE LINUX Products GmbH
Maxfeldstraße 5, 90409 Nürnberg Germany
From 67eef48cbd7a10d9451501ade7d6df9c01a6f430 Mon Sep 17 00:00:00 2001
From: Silvio Moioli smoi...@suse.de
Date: Mon, 9 Sep 2013 08:25:23 +0200
Subject: [PATCH 04/22]
AuthFilterTest was failing in our setup, apparently because the mocked
request object was missing some methods that were added after the test
itself was written. I added those methods and now it passes.
See attached patch.
Regards,
--
Silvio Moioli
SUSE LINUX Products GmbH
Maxfeldstraße 5,
OrgHandlerTest apparently required an official Red Hat channel family
with free slots:
/**
* Lookup an official Red Hat channel family with free slots.
* Fail the test if none can be found.
[...]
private ChannelFamily lookupRedHatChannelFamily() { [...] }
Actually, this is a precondition not
ActivationKeyHandlerTest expected InvalidEntitlementException where
actually FaultException gets thrown. I updated the testcase.
See attached patch.
Regards,
--
Silvio Moioli
SUSE LINUX Products GmbH
Maxfeldstraße 5, 90409 Nürnberg Germany
From 72b582c013959eb7efc45c57b05c76e7ff4bb654 Mon Sep
Hi,
I noticed that buildPageLink() uses a HashMap to keep track of query
string parameters. While this is perfectly okay from a web application
point of view, it makes testing a little more difficult and in
particular, RequestContextTest could break because it implicitly relies
on key ordering.
13 matches
Mail list logo