Merge lp:~salvatore-orlando/neutron/bug834012 into lp:neutron/diablo
Status: | Work in progress | ||||||||
---|---|---|---|---|---|---|---|---|---|
Proposed branch: | lp:~salvatore-orlando/neutron/bug834012 | ||||||||
Merge into: | lp:neutron/diablo | ||||||||
Diff against target: |
78 lines (+19/-16) 3 files modified
quantum/api/networks.py (+10/-7) quantum/api/views/networks.py (+3/-3) quantum/plugins/SamplePlugin.py (+6/-6) |
||||||||
To merge this branch: | bzr merge lp:~salvatore-orlando/neutron/bug834012 | ||||||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
dan wendlandt | Needs Fixing | ||
Review via email: mp+73687@code.launchpad.net |
Description of the change
This branch targets two bugs:
Bug #834012: API: GET networks/
Bug #837535: plugins do not comply with quantum_
They are both addressed by this branch as they are strictly related.
For Bug #834012 - changed the _item method code in the network API controller in order to avoid unnecessary calls to the plugin. The view builder was also adapted for taking into account this change.
For Bug #837535 - simply updated the documentation of quantum plugin base to state that get_network_details should return an attribute called net-ports with a sequence of ports. This seems a reasonable thing to do, as this is the way in which the currently available plugins work. Morever, and this is just my personal opinion, net-ports makes more sense than net-iface (which return the interface list only).
I realize a change in the plugin interface will probably require a deeper discussion. Please feel free to disapprove this branch if you don't agree with the proposed change.
Also, you'll see a small change in SamplePlugin. This is because it was not 100% compliant with the plugin interface.
Unmerged revisions
- 65. By Salvatore Orlando
-
Addressing Dan's comment
- 64. By Salvatore Orlando
-
Fuxing bug #834012 only
- 63. By Salvatore Orlando
-
Now properly doing the commit.
Hi Salvatore, thanks for taking care of this. I definitely agree with the goal of the patch, my only concern about timing. In general, since it seems like none of these bugs impair the correct operation of quantum, my bias would be to avoid making this change unless we have high confidence that we're not breaking plugins.
I downloaded the branch and ran unit tests with the OVS plugin and they seemed to pass. Is your feeling that the same would be true for other plugins?
Given that not all plugins have to be part of the Quantum repo, I'm still a bit wary of making a change so late, so at the very least we'd need to send an email to the list describing how plugins would have to change (assuming they were written to spec) and see if anyone raises a concern.