Merge lp:~matsubara/maas/invalid-syntax-fix into lp:~maas-committers/maas/trunk

[Diogo Matsubara]

fix invalid syntax in test_kernel_opts.py

Fix invalid syntax in test_kernel_opts.py introduce in the previous revision.

[Scott Moser (smoser)]

Thank you

review: Approve

[John A Meinel (jameinel)]

How did it get landed if it had a syntax error? nose was just not loading the file and not failing the test suite? Is there a way to teach nose that it should treat syntax errors as failures?

[Gavin Panella (allenap)]

On 19 Sep 2012 06:57, "John A Meinel" <email address hidden> wrote:
> How did it get landed if it had a syntax error? nose was just not loading
the file and not failing the test suite? Is there a way to teach nose that
it should treat syntax errors as failures?

Tests were failing, but landing to trunk is not gated on the test suite
passing; Jenkins is set up to only tests things once they're in.

Tarmac could do this for us, but we'd need to move it to a different host,
or otherwise isolate it. IIRC, it's running on carob, under Julian's
account.

Or, we could follow the Launchpad model and have a stable branch which
feeds from trunk via Jenkins.

[Julian Edwards (julian-edwards)]

On Wednesday 19 September 2012 06:19:41 you wrote:
> On 19 Sep 2012 06:57, "John A Meinel" <email address hidden> wrote:
> > How did it get landed if it had a syntax error? nose was just not loading
>
> the file and not failing the test suite? Is there a way to teach nose that
> it should treat syntax errors as failures?
>
> Tests were failing, but landing to trunk is not gated on the test suite
> passing; Jenkins is set up to only tests things once they're in.
>
> Tarmac could do this for us, but we'd need to move it to a different host,
> or otherwise isolate it. IIRC, it's running on carob, under Julian's
> account.

It is. I want to use canonistack to run it, but that is currently too
unreliable and there's no way I am running arbitrary code under my account on
carob.

> Or, we could follow the Launchpad model and have a stable branch which
> feeds from trunk via Jenkins.

This is a joke? :)

[Gavin Panella (allenap)]

On 19 September 2012 11:01, Julian Edwards <email address hidden> wrote:
>> Or, we could follow the Launchpad model and have a stable branch which
>> feeds from trunk via Jenkins.
>
> This is a joke? :)

Only half :) For the sake of development in a smallish team I think
it's good to land quickly and work on honour. For the sake of packages
it's good to build from something known to be reproducibly tested.

=== modified file 'src/provisioningserver/tests/test_kernel_opts.py'
--- src/provisioningserver/tests/test_kernel_opts.py	2012-09-18 20:48:08 +0000
+++ src/provisioningserver/tests/test_kernel_opts.py	2012-09-18 22:55:22 +0000
@@ -125,7 +125,7 @@
                 "root=LABEL=cloudimg-rootfs",
                 "iscsi_initiator=",
                 "overlayroot=tmpfs",
-                "ip=dhcp")
+                "ip=dhcp"]))
 
     def test_compose_kernel_command_line_inc_common_opts(self):
         # Test that some kernel arguments appear on both commissioning