On 03/08/2013 01:21 PM, Nishanth Aravamudan wrote:
On 08.03.2013 [11:47:36 -0300], Lucas Meneghel Rodrigues wrote:
On Thu, Mar 7, 2013 at 1:53 PM, Nishanth Aravamudan
<[email protected]> wrote:
On 27.02.2013 [22:52:19 -0300], Lucas Meneghel Rodrigues wrote:
On 02/27/2013 05:54 PM, Lucas Meneghel Rodrigues wrote:
Great, looks good.
Looks good, but the unittests will have to be adapted:
This at least makes things pass, but I think I need to still add some
tests that actual test that the username being passed in works. I would
appreciate any help with that.
I'll take care of that on a followup patch. Now, I've found another problem:
+ self.parser.add_option('-N', '--username',
+ help='Specify the username to'
+ 'login with',
+ action='store', type='string',
+ dest='username', default=None)
^ Here the shorthand -N is used for other action, so there is a
conflict when running the unittests again. I've chosen another
shorthand that doesn't conflict with others (-Q) and fixed the commit
message.
So, 2 things here:
1) Remember to run the entire unittest suite to make sure there are no
side effects to your patches:
utils/unittest_suite.py --full
Ah sorry, will do that from now on. Are the unittests safe to run in a
production instance (guarantee of no data corruption or manipulation?)
Yes, they are. The more orthodox procedure would be to transfer the
patches to your development laptop and run the tests there. Well, if it
is too much hassle (the unittests rely on things like django installed),
I suppose it's more handy to run in the prod instance.
I personally keep my laptop able to run the unittests, with a subset of
the packages that you can install with
installation_support/autotest-install-packages-deps.
_______________________________________________
Autotest-kernel mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/autotest-kernel
