)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"d8050f19e1d50b32a956145dca20cbb155f889c1","unresolved":false,"context_lines":[{"line_number":12,"context_line":"of \u0027br-int\u0027) is created [2]."},{"line_number":13,"context_line":""},{"line_number":14,"context_line":"Before an OVS port\u0027s vif_details are committed, the openvswitch driver"},{"line_number":15,"context_line":"for the trunk service plugin will check if the port is a trunk and"},{"line_number":16,"context_line":"generate a trunk bridge name for the vif_detail\u0027s \u0027bridge_name\u0027."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Note: until os-vif has been integrated into Nova [3], the trunk bridge"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dada55a8_d82cd0ad","line":15,"updated":"2016-07-18 17:42:57.000000000","message":"Can we limit this commit message to simply describe that this patch can be used by the vlan-aware-vms work, and avoid getting into implementation details?","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"d4f938c308f9975482f5b18499a4b9892955521a","unresolved":false,"context_lines":[{"line_number":25,"context_line":"[2] https://review.openstack.org/#/c/330818/"},{"line_number":26,"context_line":"[3] https://review.openstack.org/#/c/269672/"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Partially-implements: blueprint vlan-aware-vms"},{"line_number":29,"context_line":"Change-Id: Iad72d163bc406df65866f6d962abcb78596828e7"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dada55a8_f8e554a8","line":28,"updated":"2016-07-18 17:38:37.000000000","message":"This is really https://blueprints.launchpad.net/nova/+spec/neutron-ovs-bridge-name\n\nNothing to do with vlan-aware-vms per se, though this work is needed to enable the trunk use case.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"f6ad7275db12ac735095bc9e6c8a1f9010f1c5a8","unresolved":false,"context_lines":[{"line_number":25,"context_line":"[2] https://review.openstack.org/#/c/330818/"},{"line_number":26,"context_line":"[3] https://review.openstack.org/#/c/269672/"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Partially-implements: blueprint vlan-aware-vms"},{"line_number":29,"context_line":"Change-Id: Iad72d163bc406df65866f6d962abcb78596828e7"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dada55a8_b9aabdd2","line":28,"in_reply_to":"dada55a8_789fa403","updated":"2016-07-18 20:09:10.000000000","message":"Ok, it seems like you\u0027re saying this should be split into two patches.\nPatch 1 enables a generic way for the OVS mech_driver to set bridge_name in the vif_details and expose it for mutation to other interested parties. That patch will target the neutron-ovs-bridge-name blueprint and will contain the changes in this patch except for the trunk ovs driver change.\nPatch 2 has the trunk OVS driver change that subscribes the callback that mutates the bridge_name and targets vlan-aware-vms.\n\nIs that what you\u0027re getting at?","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"e5faffb0d2e2c90b3ba39e5a837cf72473911908","unresolved":false,"context_lines":[{"line_number":25,"context_line":"[2] https://review.openstack.org/#/c/330818/"},{"line_number":26,"context_line":"[3] https://review.openstack.org/#/c/269672/"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Partially-implements: blueprint vlan-aware-vms"},{"line_number":29,"context_line":"Change-Id: Iad72d163bc406df65866f6d962abcb78596828e7"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dada55a8_bc71d9e2","line":28,"in_reply_to":"dada55a8_b9aabdd2","updated":"2016-07-19 16:36:07.000000000","message":"I am saying that exposing the bridge details is not something that must be tied to vlan-aware-vms, it\u0027s a piece of work in its own right.\n\nNow we can split it in two or keep it as one, it depends on how bulky the patch becomes and how logical the split is. See my next comment, if we agree that\u0027s a sensible approach, then we can definitely split the patch in two, one that targets neutron-ovs-bridge-name the other that targets vlan-aware-vms.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"d8050f19e1d50b32a956145dca20cbb155f889c1","unresolved":false,"context_lines":[{"line_number":25,"context_line":"[2] https://review.openstack.org/#/c/330818/"},{"line_number":26,"context_line":"[3] https://review.openstack.org/#/c/269672/"},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Partially-implements: blueprint vlan-aware-vms"},{"line_number":29,"context_line":"Change-Id: Iad72d163bc406df65866f6d962abcb78596828e7"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"dada55a8_789fa403","line":28,"in_reply_to":"dada55a8_f8e554a8","updated":"2016-07-18 17:42:57.000000000","message":"I\u0027d target the other one or both at least.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"}],"neutron/plugins/ml2/drivers/openvswitch/mech_driver/mech_openvswitch.py":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"e5faffb0d2e2c90b3ba39e5a837cf72473911908","unresolved":false,"context_lines":[{"line_number":73,"context_line":"    def try_to_bind_segment_for_agent(self, context, segment, agent):"},{"line_number":74,"context_line":"        if self.check_segment_for_agent(segment, agent):"},{"line_number":75,"context_line":"            vif_details \u003d self.get_vif_details(agent, context)"},{"line_number":76,"context_line":"            registry.notify("},{"line_number":77,"context_line":"                portbindings.OVS_VIF_DETAILS, events.PRECOMMIT_CREATE, self,"},{"line_number":78,"context_line":"                **{\u0027port\u0027: context.current,"},{"line_number":79,"context_line":"                   \u0027vif_details\u0027: vif_details})"}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_c28784d5","line":76,"updated":"2016-07-19 16:36:07.000000000","message":"I was thinking that to implement the blueprint neutron-ovs-bridge-name we would add a new model Port\u003c-\u003eBridgeName. This captures the relationship between the port and the bridge (it can either be implemented as its own table or a NULL column added to the Port model).\n\nOnce that\u0027s done, the method get_vif_details would include bridge_name in the vif_details if the value is set for the port_id. Now it\u0027s the job of someone else to set the bridge for the port. In the case of vlan aware vms, the value can be set when we create a trunk, as at that point we know the parent port and can generate a bridge name and associate it to the port itself.\n\nDuring the subsequent binding process, the bridge_name is reflected in the vif details.\n\nWe could potentially get away without a model addition, but it would make the existing vif_details handling code prone to errors as the binding may get wiped out depending on the state the port is in.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"7bc744746bcb6c3b1b22d90d8c246979167c6faa","unresolved":false,"context_lines":[{"line_number":73,"context_line":"    def try_to_bind_segment_for_agent(self, context, segment, agent):"},{"line_number":74,"context_line":"        if self.check_segment_for_agent(segment, agent):"},{"line_number":75,"context_line":"            vif_details \u003d self.get_vif_details(agent, context)"},{"line_number":76,"context_line":"            registry.notify("},{"line_number":77,"context_line":"                portbindings.OVS_VIF_DETAILS, events.PRECOMMIT_CREATE, self,"},{"line_number":78,"context_line":"                **{\u0027port\u0027: context.current,"},{"line_number":79,"context_line":"                   \u0027vif_details\u0027: vif_details})"}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_702ad42d","line":76,"in_reply_to":"dada55a8_2abe29d4","updated":"2016-07-19 22:13:35.000000000","message":"For point C, the ovs trunk driver could generate a trunk bridge_name for the port when it\u0027s used to create a trunk. But, what if the linuxbridge trunk driver (or some other driver) wants to set the bridge_name too? Before the trunk port is actually bound, we don\u0027t know which driver will actually \"own\" that trunk port.\n\nTo me the bridge_name is basically just to tell Nova to \"plug the vif into a bridge with this name\". What other use cases would you have in mind for bridge_name?\n\nI think we could be setting the bridge_name unconditionally in the mech_drivers to a sane default, i.e. cfg.CONF.OVS.integration_bridge for OVS, and \u0027brq\u0027+port[\u0027network_id\u0027] for LinuxBridge. Then special use cases such as vlan-aware-vms could change the bridge_name to suit their vif-plugging needs. I\u0027m not sure what other use cases might require plugging the vif into a special bridge other than \u0027br-int\u0027 besides vlan-aware-vms. If another use case arises that requires its own special bridge, then the two use cases are probably incompatible, and we should only allow one mutator function for bridge_name.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"0b1654050c3b7cc0bbd0e47c328ee0dad3503fb9","unresolved":false,"context_lines":[{"line_number":73,"context_line":"    def try_to_bind_segment_for_agent(self, context, segment, agent):"},{"line_number":74,"context_line":"        if self.check_segment_for_agent(segment, agent):"},{"line_number":75,"context_line":"            vif_details \u003d self.get_vif_details(agent, context)"},{"line_number":76,"context_line":"            registry.notify("},{"line_number":77,"context_line":"                portbindings.OVS_VIF_DETAILS, events.PRECOMMIT_CREATE, self,"},{"line_number":78,"context_line":"                **{\u0027port\u0027: context.current,"},{"line_number":79,"context_line":"                   \u0027vif_details\u0027: vif_details})"}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_30cb3c20","line":76,"in_reply_to":"dada55a8_702ad42d","updated":"2016-07-19 23:24:16.000000000","message":"You\u0027re assuming that driver X (agent or agentless it may be) may need to use the bridge_name in a trunk application and in the context of the same deployment. Is it a reasonable assumption? Now let\u0027s assume that OVS, LinuxBridge, OVN and ODL were to co-exist in the same deployment; none of drivers that allegedly implement the trunk extension would make use of the bridge_name except OVS (I am not 100% sure about ODL but I doubt it),  so the conflict is purely hypothetical at this point. Having said that, should deferring the bridge generation be absolutely necessary in the future, the idea of generating the bridge name at the time of the binding process could still be pursued but I wonder if doing it at the time of the caller (i.e. bind_port, and thus explicitly rather than implicitly as prescribed in [1,2]), is more appropriate. Requirement [1,2] has always smelt funny to me because it spills implementation details and cross-cutting concerns, but that\u0027s a separate discussion.\n\nAs for the use of the bridge name, yes, the only use case from a Nova point of view is \u0027plug this port here\u0027. The use cases may vary depending on what that bridge is.\n\nIrrespective of what strategy we choose, I believe that the mech driver is in charge of setting up the bridge_name explicitly rather than as a side effect as done at line 79. How that bridge_name is determined may either be done ahead of time or at bind time, and if we go with your proposal we may want to come up with a more defined contract to determine this information, e.g. we can leverage/expose an API that, given the agent type, vnic_type - and whatever else needed - can determine the bridge_name. The provider of this API may as well be the mech_agent class that \u0027asks\u0027 for the bridge to be set before vif_details are propagated to the binding. A bit like you did here but slightly more formally.\n\n[1] https://github.com/openstack/neutron/blob/master/neutron/plugins/ml2/driver_api.py#L864\n[2] https://github.com/openstack/neutron/blob/master/neutron/plugins/ml2/drivers/mech_agent.py#L132","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"1285f267aa926311fca8a79fc220464a7286de85","unresolved":false,"context_lines":[{"line_number":73,"context_line":"    def try_to_bind_segment_for_agent(self, context, segment, agent):"},{"line_number":74,"context_line":"        if self.check_segment_for_agent(segment, agent):"},{"line_number":75,"context_line":"            vif_details \u003d self.get_vif_details(agent, context)"},{"line_number":76,"context_line":"            registry.notify("},{"line_number":77,"context_line":"                portbindings.OVS_VIF_DETAILS, events.PRECOMMIT_CREATE, self,"},{"line_number":78,"context_line":"                **{\u0027port\u0027: context.current,"},{"line_number":79,"context_line":"                   \u0027vif_details\u0027: vif_details})"}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_cf5f9f3f","line":76,"in_reply_to":"dada55a8_c28784d5","updated":"2016-07-19 21:01:42.000000000","message":"I think a model addition is overkill here because the bridge_name has different meaning depending on the portbinding vif_type (\u0027ovs\u0027 or \u0027bridge\u0027 for instance), and that is known only once the port is bound. Losing the bridge_name when the port becomes unbound is OK, because it only makes sense in the context of a bound port. When the port is unbound, we no longer know if it\u0027s going to end up being an OVS port or not, so it makes sense to wipe out the bridge_name.\n\nFor vlan-aware-vms, we cannot set bridge_name on a port when the trunk is created, because we don\u0027t know what vif_type the trunk port will be bound with yet. Say we set the bridge_name on that port to the generated OVS trunk bridge name, but then the trunk becomes bound as vif_type\u003d\u003d\u0027bridge\u0027 (for LinuxBridge). That won\u0027t work.\n\nThat\u0027s why I think this belongs to the individual mech_drivers that generate the vif_details that bridge_name will be part of.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"68c23aa11bb2a1a0161bdd7aa4dabd82c267ac84","unresolved":false,"context_lines":[{"line_number":73,"context_line":"    def try_to_bind_segment_for_agent(self, context, segment, agent):"},{"line_number":74,"context_line":"        if self.check_segment_for_agent(segment, agent):"},{"line_number":75,"context_line":"            vif_details \u003d self.get_vif_details(agent, context)"},{"line_number":76,"context_line":"            registry.notify("},{"line_number":77,"context_line":"                portbindings.OVS_VIF_DETAILS, events.PRECOMMIT_CREATE, self,"},{"line_number":78,"context_line":"                **{\u0027port\u0027: context.current,"},{"line_number":79,"context_line":"                   \u0027vif_details\u0027: vif_details})"}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_2abe29d4","line":76,"in_reply_to":"dada55a8_cf5f9f3f","updated":"2016-07-19 21:25:50.000000000","message":"Ok, you brought out a few points:\n\na) The bridge_name is overkill: the fact that bridge_name may have different meanings has nothing to do with whether model is needed or not. That\u0027s why I said that bridge_name be a nullable property, and if we go for its own table, the association record will be there only when needed. Besides, I disagree that the bridge name may end up being known only when the port is bound. There are clear circumstances like trunks where the bridge is well known in advance, especially when a port is selected to be a parent in trunk.\n\nb) Cleaning up the bridge association is something that can be taken care of once we define what rules must be followed, that is not a point of contention (unless I misunderstood what you\u0027re implying), so I am not worried.\n\nc) We cannot generate the bridge name ahead of time because that may lead to conflicts: I am not suggesting we create this unconditionally.  The bridge name generation will have to happen by the ovs trunk driver and would be picked up by the mech_openvswitch driver. If other drivers know nothing about it, they won\u0027t know about it because it might not even be there. So I wouldn\u0027t be hasty in claiming it won\u0027t work. \n\nd) The generation logic belong to the individual mech_drivers: I believe this is wrong. You\u0027re coupling trunk details with the OVS mech driver when in reality the bridge name may have different applications and it would be difficult to adopt different use cases the way it\u0027s done in this first patchset. By simply making get_vif_details pass the bridge_name if present, we can address multiple application needs without affecting the OVS driver every time.\n\nAfter going through your response I am even more convinced that the approach I am proposing is a reasonable one, and I would urge to reconsider. That said, I am happy to see what others might be thinking.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":4187,"name":"Ryan Tidwell","email":"rtidwell@suse.com","username":"ryan-tidwell"},"change_message_id":"8959c41e5245502a1f9997a1cc1b010c6f3271b6","unresolved":false,"context_lines":[{"line_number":98,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":99,"context_line":"            vif_details[portbindings.VIF_DETAILS_BRIDGE_NAME] \u003d bridge_name"},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"        registry.notify("},{"line_number":102,"context_line":"            a_const.OVS_BRIDGE_NAME, events.PRECOMMIT_CREATE,"},{"line_number":103,"context_line":"            set_bridge_name_inner, port\u003dport)"},{"line_number":104,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"dada55a8_98dc8fe5","line":101,"updated":"2016-07-26 21:40:36.000000000","message":"Not an issue *yet*, but what happens when you have multiple subscribers to this event and set different bridge names? Doesn\u0027t it becomes a situation where the last callback invoked wins? This is extremely de-coupled, but it does feel  little weird that I could subscribe to this event and override the bridge name set by other drivers, and they could override my bridge name?","commit_id":"1cf77b0cef2c12fa66e6c1f076b319633b18bc88"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2180c4a9c907063f9153d72426edd64bc28db7d1","unresolved":false,"context_lines":[{"line_number":98,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":99,"context_line":"            vif_details[portbindings.VIF_DETAILS_BRIDGE_NAME] \u003d bridge_name"},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"        registry.notify("},{"line_number":102,"context_line":"            a_const.OVS_BRIDGE_NAME, events.PRECOMMIT_CREATE,"},{"line_number":103,"context_line":"            set_bridge_name_inner, port\u003dport)"},{"line_number":104,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"dada55a8_c0d8c920","line":101,"in_reply_to":"dada55a8_2ff556ec","updated":"2016-07-27 00:13:42.000000000","message":"That\u0027s what I implied in [1], and I agree that we could prevent interference by preventing setting the bridge more than once.\n\n[1] https://review.openstack.org/#/c/343816/2/neutron/services/trunk/drivers/openvswitch.py@48","commit_id":"1cf77b0cef2c12fa66e6c1f076b319633b18bc88"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"f0f9689daa1de9fdd9ba5a9c6ead3083ea20a7f7","unresolved":false,"context_lines":[{"line_number":98,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":99,"context_line":"            vif_details[portbindings.VIF_DETAILS_BRIDGE_NAME] \u003d bridge_name"},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"        registry.notify("},{"line_number":102,"context_line":"            a_const.OVS_BRIDGE_NAME, events.PRECOMMIT_CREATE,"},{"line_number":103,"context_line":"            set_bridge_name_inner, port\u003dport)"},{"line_number":104,"context_line":""}],"source_content_type":"text/x-python","patch_set":3,"id":"dada55a8_2ff556ec","line":101,"in_reply_to":"dada55a8_98dc8fe5","updated":"2016-07-26 22:22:10.000000000","message":"Yeah, I had a comment on an earlier patchset that mentioned this. Should we prohibit set_bridge_name from being called more than once and throw an exception in that case? If two subscribers both need to set the bridge_name, I feel like the incompatibility of the two drivers/plugins/whatever should be handled formally at a much higher level, but I\u0027m ok with going the extra mile and enforcing this from the notifier.","commit_id":"1cf77b0cef2c12fa66e6c1f076b319633b18bc88"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"b36caac9cf018f340a8e38b700234be192bc95ff","unresolved":false,"context_lines":[{"line_number":96,"context_line":"    def _set_bridge_name(self, port, vif_details):"},{"line_number":97,"context_line":"        self._bridge_name_set \u003d False"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":100,"context_line":"            if self._bridge_name_set:"},{"line_number":101,"context_line":"                raise RuntimeError(_(\"An OVS port\u0027s bridge_name cannot be \""},{"line_number":102,"context_line":"                                     \"set more than once.\"))"}],"source_content_type":"text/x-python","patch_set":18,"id":"bacf61ea_dd6b84ef","line":99,"updated":"2016-08-05 16:44:58.000000000","message":"This is not wrong, but I think it could have been done more cleanly. To avoid multiple attempts to return a bridge name I would have solved it differently: I personally would have provided an API for interested parties to be able to advertise the bridge name. The API in turn would have registered the callback and fail in case someone was attempting to register a duplicate callback. The registration happens much earlier in the process and therefore we would have failed fast, and avoided the RunTime error during the binding process, which is usually a pain in the rear end to troubleshoot.","commit_id":"9ee57e85a2ca9013cba643bda0d2613585a97129"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"521ee42c78bf44eed7245713dc271fa1420ede77","unresolved":false,"context_lines":[{"line_number":96,"context_line":"    def _set_bridge_name(self, port, vif_details):"},{"line_number":97,"context_line":"        self._bridge_name_set \u003d False"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":100,"context_line":"            if self._bridge_name_set:"},{"line_number":101,"context_line":"                raise RuntimeError(_(\"An OVS port\u0027s bridge_name cannot be \""},{"line_number":102,"context_line":"                                     \"set more than once.\"))"}],"source_content_type":"text/x-python","patch_set":18,"id":"bacf61ea_32048613","line":99,"in_reply_to":"bacf61ea_3bd468ea","updated":"2016-08-05 18:39:25.000000000","message":"One more thing: as for the order, there\u0027s no ordering issue here: the notify happens at runtime (when you bind ports) and that happens only after the server is up and running (where the subscribes take place). Not sure what you mean by AFTER_INIT event as I see none.","commit_id":"9ee57e85a2ca9013cba643bda0d2613585a97129"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"1d540ff65b5185596c9ee0e6586d0a5258b83c43","unresolved":false,"context_lines":[{"line_number":96,"context_line":"    def _set_bridge_name(self, port, vif_details):"},{"line_number":97,"context_line":"        self._bridge_name_set \u003d False"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":100,"context_line":"            if self._bridge_name_set:"},{"line_number":101,"context_line":"                raise RuntimeError(_(\"An OVS port\u0027s bridge_name cannot be \""},{"line_number":102,"context_line":"                                     \"set more than once.\"))"}],"source_content_type":"text/x-python","patch_set":18,"id":"bacf61ea_d1b7eb82","line":99,"in_reply_to":"bacf61ea_3bd468ea","updated":"2016-08-05 18:08:09.000000000","message":"The order is guaranteed, and core setup comes first. That said we can revise this later once the dust settles. We\u0027d need to enhance the registry to accommodate my idea.","commit_id":"9ee57e85a2ca9013cba643bda0d2613585a97129"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"ca446df15b0d0b4064fdc490e465c67ccc505bd1","unresolved":false,"context_lines":[{"line_number":96,"context_line":"    def _set_bridge_name(self, port, vif_details):"},{"line_number":97,"context_line":"        self._bridge_name_set \u003d False"},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":100,"context_line":"            if self._bridge_name_set:"},{"line_number":101,"context_line":"                raise RuntimeError(_(\"An OVS port\u0027s bridge_name cannot be \""},{"line_number":102,"context_line":"                                     \"set more than once.\"))"}],"source_content_type":"text/x-python","patch_set":18,"id":"bacf61ea_3bd468ea","line":99,"in_reply_to":"bacf61ea_dd6b84ef","updated":"2016-08-05 17:33:19.000000000","message":"I think I grasp what you mean, but what I\u0027m envisioning is predicated on the trunk drivers being initialized before the ovs mech_driver so that they can subscribe callbacks to the mech_driver\u0027s AFTER_INIT notification. I\u0027m not sure what order they\u0027re initialized upon startup, but assuming a certain order seems fragile. There should be a way to do it that doesn\u0027t assume a certain order, but I\u0027m not inclined to go down that path unless we\u0027re wholly unsatisfied with this approach. Really the error should only ever be hit by developers of new drivers, which if they\u0027ve figured out how to subscribe to this notification, they should already realize that the bridge_name shouldn\u0027t be set more than once.","commit_id":"9ee57e85a2ca9013cba643bda0d2613585a97129"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"6049ede98e017ec13e153a7a1f740690807ca327","unresolved":false,"context_lines":[{"line_number":99,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":100,"context_line":"            if self._bridge_name_set:"},{"line_number":101,"context_line":"                raise RuntimeError(_(\"An OVS port\u0027s bridge_name cannot be \""},{"line_number":102,"context_line":"                                     \"set more than once.\"))"},{"line_number":103,"context_line":"            else:"},{"line_number":104,"context_line":"                vif_details[portbindings.VIF_DETAILS_BRIDGE_NAME] \u003d bridge_name"},{"line_number":105,"context_line":"                self._bridge_name_set \u003d True"}],"source_content_type":"text/x-python","patch_set":19,"id":"bacf61ea_fb4e4936","line":102,"updated":"2016-08-05 20:15:21.000000000","message":"I am having second thoughts about this, I think for now we should not worry about multiple callbacks registering for this event. I don\u0027t quite like as it feels too coupled, it and I\u0027d rather have this not dealt with, than dealt with this way.\n\nI would like this cleaned up anyway as I envisioned in PS1 where a bridge_name for the port is introduced explicitly in the schema.\n\n[1] https://review.openstack.org/#/c/343816/1/neutron/plugins/ml2/drivers/openvswitch/mech_driver/mech_openvswitch.py@76","commit_id":"cf5cd7cc7ee2e21a6db43ecea3529bdd3b060d69"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"dfd6f01bb2fec7935b9c774fe8375faedf6601fe","unresolved":false,"context_lines":[{"line_number":99,"context_line":"        def set_bridge_name_inner(bridge_name):"},{"line_number":100,"context_line":"            if self._bridge_name_set:"},{"line_number":101,"context_line":"                raise RuntimeError(_(\"An OVS port\u0027s bridge_name cannot be \""},{"line_number":102,"context_line":"                                     \"set more than once.\"))"},{"line_number":103,"context_line":"            else:"},{"line_number":104,"context_line":"                vif_details[portbindings.VIF_DETAILS_BRIDGE_NAME] \u003d bridge_name"},{"line_number":105,"context_line":"                self._bridge_name_set \u003d True"}],"source_content_type":"text/x-python","patch_set":19,"id":"9ad45d7e_462650bb","line":102,"in_reply_to":"bacf61ea_fb4e4936","updated":"2016-08-08 18:56:12.000000000","message":"I removed the check against multiple callbacks registering for this event and added a comment to revisit your proposed approach.","commit_id":"cf5cd7cc7ee2e21a6db43ecea3529bdd3b060d69"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"6049ede98e017ec13e153a7a1f740690807ca327","unresolved":false,"context_lines":[{"line_number":105,"context_line":"                self._bridge_name_set \u003d True"},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"        registry.notify("},{"line_number":108,"context_line":"            a_const.OVS_BRIDGE_NAME, events.PRECOMMIT_CREATE,"},{"line_number":109,"context_line":"            set_bridge_name_inner, port\u003dport)"},{"line_number":110,"context_line":""},{"line_number":111,"context_line":"    def _pre_get_vif_details(self, agent, context):"}],"source_content_type":"text/x-python","patch_set":19,"id":"bacf61ea_3ba0e1f3","line":108,"updated":"2016-08-05 20:15:21.000000000","message":"BTW, I am thinking that perhaps BEFORE_READ is more appropriate.","commit_id":"cf5cd7cc7ee2e21a6db43ecea3529bdd3b060d69"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"dfd6f01bb2fec7935b9c774fe8375faedf6601fe","unresolved":false,"context_lines":[{"line_number":105,"context_line":"                self._bridge_name_set \u003d True"},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"        registry.notify("},{"line_number":108,"context_line":"            a_const.OVS_BRIDGE_NAME, events.PRECOMMIT_CREATE,"},{"line_number":109,"context_line":"            set_bridge_name_inner, port\u003dport)"},{"line_number":110,"context_line":""},{"line_number":111,"context_line":"    def _pre_get_vif_details(self, agent, context):"}],"source_content_type":"text/x-python","patch_set":19,"id":"9ad45d7e_262b5c84","line":108,"in_reply_to":"bacf61ea_3ba0e1f3","updated":"2016-08-08 18:56:12.000000000","message":"Done","commit_id":"cf5cd7cc7ee2e21a6db43ecea3529bdd3b060d69"}],"neutron/services/trunk/drivers/openvswitch.py":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"d8050f19e1d50b32a956145dca20cbb155f889c1","unresolved":false,"context_lines":[{"line_number":41,"context_line":"def vif_details_handler(resource, event, trigger, **kwargs):"},{"line_number":42,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":43,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":44,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":45,"context_line":"        vif_details \u003d kwargs[\u0027vif_details\u0027]"},{"line_number":46,"context_line":"        vif_details[pb.VIF_DETAILS_BRIDGE_NAME] \u003d gen_trunk_br_name(port[\u0027id\u0027])"},{"line_number":47,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_b33adfcb","line":44,"updated":"2016-07-18 17:42:57.000000000","message":"we should making this about vlan aware vms. We should come up with an approach that will let the OVS agent pass the bridge name in such a way that it doesn\u0027t not fulfil a single use case, namely the vlan-aware-vms one.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"f6ad7275db12ac735095bc9e6c8a1f9010f1c5a8","unresolved":false,"context_lines":[{"line_number":41,"context_line":"def vif_details_handler(resource, event, trigger, **kwargs):"},{"line_number":42,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":43,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":44,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":45,"context_line":"        vif_details \u003d kwargs[\u0027vif_details\u0027]"},{"line_number":46,"context_line":"        vif_details[pb.VIF_DETAILS_BRIDGE_NAME] \u003d gen_trunk_br_name(port[\u0027id\u0027])"},{"line_number":47,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_f92275b8","line":44,"in_reply_to":"dada55a8_b33adfcb","updated":"2016-07-18 20:09:10.000000000","message":"So, the problem with this is that if multiple parties want to modify the bridge_name, they will overwrite each other. Is that right?\n\nWhat if we protected against overwrite by providing a generic way for the OVS mech_driver to allow a single method to update a single portion of the vif_details, similar to how we register a function for validating segmentation_ids per segmentation_type (currently just vlan). If more than one function is registered to modify the bridge_name, the ovs mech_driver will throw an exception and disallow it.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"e5faffb0d2e2c90b3ba39e5a837cf72473911908","unresolved":false,"context_lines":[{"line_number":41,"context_line":"def vif_details_handler(resource, event, trigger, **kwargs):"},{"line_number":42,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":43,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":44,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":45,"context_line":"        vif_details \u003d kwargs[\u0027vif_details\u0027]"},{"line_number":46,"context_line":"        vif_details[pb.VIF_DETAILS_BRIDGE_NAME] \u003d gen_trunk_br_name(port[\u0027id\u0027])"},{"line_number":47,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"dada55a8_f6cfed7c","line":44,"in_reply_to":"dada55a8_f92275b8","updated":"2016-07-19 16:36:07.000000000","message":"See if my comment to the previous file makes sense. If we agree that\u0027s a viable integration strategy, then we don\u0027t need the callbacks at all.\n\nA point worth making on your callback usage here is that you\u0027re not supposed to cause side effects on the payload (e.g. these are meant to be passed by copy rather than by value). Here you change the vif details on the fly and if there are multiple callbacks that may lead to interference.\n\nA more predictable way to make callback and notify interact would be to use message passing via the trigger object. For instance, you could call a method on the trigger to pass information back to it.","commit_id":"e26b979eae7445f0ea22b795ec9d6689d76b6f9a"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"31607067795a58dfc46b8f00d034d383e30d4b31","unresolved":false,"context_lines":[{"line_number":22,"context_line":"from neutron.callbacks import registry"},{"line_number":23,"context_line":"from neutron.common import rpc as n_rpc"},{"line_number":24,"context_line":"from neutron import manager"},{"line_number":25,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.agent.common import constants \\"},{"line_number":26,"context_line":"    as agent_consts"},{"line_number":27,"context_line":"from neutron.services.trunk import constants as trunk_consts"},{"line_number":28,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_93f4adcf","line":25,"updated":"2016-07-26 00:15:25.000000000","message":"if you respin, please use ( rather than \\","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"6cbd74dcbcc78a20656f8ea9da3acb71cbbc5e15","unresolved":false,"context_lines":[{"line_number":22,"context_line":"from neutron.callbacks import registry"},{"line_number":23,"context_line":"from neutron.common import rpc as n_rpc"},{"line_number":24,"context_line":"from neutron import manager"},{"line_number":25,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.agent.common import constants \\"},{"line_number":26,"context_line":"    as agent_consts"},{"line_number":27,"context_line":"from neutron.services.trunk import constants as trunk_consts"},{"line_number":28,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_14523967","line":25,"in_reply_to":"dada55a8_93f4adcf","updated":"2016-07-26 16:52:34.000000000","message":"Done","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"31607067795a58dfc46b8f00d034d383e30d4b31","unresolved":false,"context_lines":[{"line_number":43,"context_line":""},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"def vif_details_bridge_name_handler(resource, event, trigger, **kwargs):"},{"line_number":46,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":47,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":48,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_332ce149","line":46,"updated":"2016-07-26 00:15:25.000000000","message":"missing trailing period.","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"6cbd74dcbcc78a20656f8ea9da3acb71cbbc5e15","unresolved":false,"context_lines":[{"line_number":43,"context_line":""},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"def vif_details_bridge_name_handler(resource, event, trigger, **kwargs):"},{"line_number":46,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":47,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":48,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_d4d261d0","line":46,"in_reply_to":"dada55a8_332ce149","updated":"2016-07-26 16:52:34.000000000","message":"Done","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"31607067795a58dfc46b8f00d034d383e30d4b31","unresolved":false,"context_lines":[{"line_number":45,"context_line":"def vif_details_bridge_name_handler(resource, event, trigger, **kwargs):"},{"line_number":46,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":47,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":48,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_f3c88976","line":48,"updated":"2016-07-26 00:15:25.000000000","message":"we should add a NOTE reminding ourselves that this method could be made general and provide either ovs_integration_bridge or the trunk bridge name if the trunk extension is available, but it\u0027s a nit.","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"4ff9e0398d509946b493b68dd2fcd313bf35c43d","unresolved":false,"context_lines":[{"line_number":45,"context_line":"def vif_details_bridge_name_handler(resource, event, trigger, **kwargs):"},{"line_number":46,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":47,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":48,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_37277f13","line":48,"in_reply_to":"dada55a8_941d8975","updated":"2016-07-26 16:56:30.000000000","message":"No. We need an overarching element that takes care of the bridge setup on the mech-driver\u0027s behalf, otherwise there\u0027s a risk for interference. At that point the logic would have to move out of the responsibility of the trunk plugin.","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"6cbd74dcbcc78a20656f8ea9da3acb71cbbc5e15","unresolved":false,"context_lines":[{"line_number":45,"context_line":"def vif_details_bridge_name_handler(resource, event, trigger, **kwargs):"},{"line_number":46,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details\"\"\""},{"line_number":47,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":48,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_941d8975","line":48,"in_reply_to":"dada55a8_f3c88976","updated":"2016-07-26 16:52:34.000000000","message":"Do you think it should be the ovs trunk driver\u0027s responsibility to set it to ovs_integration_bridge if it\u0027s not a trunk port? I was thinking that possibly the ovs mech_driver would be responsible for setting that by default. Then, this callback would overwrite the default for trunk ports.","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"31607067795a58dfc46b8f00d034d383e30d4b31","unresolved":false,"context_lines":[{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"def gen_trunk_br_name(port_id):"},{"line_number":53,"context_line":"    return (\u0027tbr-\u0027 + port_id)[:lib_consts.DEVICE_NAME_MAX_LEN - 1]"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_33a181aa","line":52,"updated":"2016-07-26 00:15:25.000000000","message":"this may end up in a different place, but for now I guess there\u0027s no better place.","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"6cbd74dcbcc78a20656f8ea9da3acb71cbbc5e15","unresolved":false,"context_lines":[{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"def gen_trunk_br_name(port_id):"},{"line_number":53,"context_line":"    return (\u0027tbr-\u0027 + port_id)[:lib_consts.DEVICE_NAME_MAX_LEN - 1]"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_f71cc767","line":52,"in_reply_to":"dada55a8_33a181aa","updated":"2016-07-26 16:52:34.000000000","message":"ack, I noticed your comment on [1]. I\u0027ll add a TODO here make use of the common util method when available.\n\n[1] https://review.openstack.org/#/c/335536/","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"bc07cb2f053cbece8024b35aae4a2987bab80f56","unresolved":false,"context_lines":[{"line_number":49,"context_line":"        trigger(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"def gen_trunk_br_name(port_id):"},{"line_number":53,"context_line":"    return (\u0027tbr-\u0027 + port_id)[:lib_consts.DEVICE_NAME_MAX_LEN - 1]"},{"line_number":54,"context_line":""},{"line_number":55,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_376e5f2f","line":52,"in_reply_to":"dada55a8_f71cc767","updated":"2016-07-26 16:57:03.000000000","message":"yeah, that\u0027ll do.","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"}],"neutron/services/trunk/drivers/openvswitch/constants.py":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"b36caac9cf018f340a8e38b700234be192bc95ff","unresolved":false,"context_lines":[{"line_number":12,"context_line":"# License for the specific language governing permissions and limitations"},{"line_number":13,"context_line":"# under the License."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"TRUNK_BR_PREFIX \u003d \u0027tbr-\u0027"}],"source_content_type":"text/x-python","patch_set":18,"id":"bacf61ea_5d579425","line":15,"updated":"2016-08-05 16:44:58.000000000","message":"add a TODO so that we can reconcile with:\n\nhttps://review.openstack.org/#/c/335536/9/neutron/services/trunk/utils.py","commit_id":"9ee57e85a2ca9013cba643bda0d2613585a97129"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"ca446df15b0d0b4064fdc490e465c67ccc505bd1","unresolved":false,"context_lines":[{"line_number":12,"context_line":"# License for the specific language governing permissions and limitations"},{"line_number":13,"context_line":"# under the License."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"TRUNK_BR_PREFIX \u003d \u0027tbr-\u0027"}],"source_content_type":"text/x-python","patch_set":18,"id":"bacf61ea_36447d48","line":15,"in_reply_to":"bacf61ea_5d579425","updated":"2016-08-05 17:33:19.000000000","message":"Done","commit_id":"9ee57e85a2ca9013cba643bda0d2613585a97129"}],"neutron/services/trunk/drivers/openvswitch/driver.py":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"dd683ae283c6b30a442580c52a1823d2049a9e96","unresolved":false,"context_lines":[{"line_number":56,"context_line":"        set_br_name(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"def gen_trunk_br_name(port_id):"},{"line_number":60,"context_line":"    # TODO(rawlin): this function should be moved to a common utils module"},{"line_number":61,"context_line":"    return (\u0027tbr-\u0027 + port_id)[:constants.DEVICE_NAME_MAX_LEN - 1]"},{"line_number":62,"context_line":""}],"source_content_type":"text/x-python","patch_set":8,"id":"bacf61ea_dc04a2c1","line":59,"updated":"2016-07-28 18:30:00.000000000","message":"why don\u0027t we create one already? Like under neutron/services/trunk/utils?","commit_id":"8e924a9b2cd1042567b69c58694d385531b08a78"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"da59a9dc878984585e13afaf50ee8b6daf0fe01e","unresolved":false,"context_lines":[{"line_number":56,"context_line":"        set_br_name(gen_trunk_br_name(port[\u0027id\u0027]))"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"def gen_trunk_br_name(port_id):"},{"line_number":60,"context_line":"    # TODO(rawlin): this function should be moved to a common utils module"},{"line_number":61,"context_line":"    return (\u0027tbr-\u0027 + port_id)[:constants.DEVICE_NAME_MAX_LEN - 1]"},{"line_number":62,"context_line":""}],"source_content_type":"text/x-python","patch_set":8,"id":"bacf61ea_252162b5","line":59,"in_reply_to":"bacf61ea_dc04a2c1","updated":"2016-07-28 19:26:04.000000000","message":"Done","commit_id":"8e924a9b2cd1042567b69c58694d385531b08a78"},{"author":{"_account_id":4187,"name":"Ryan Tidwell","email":"rtidwell@suse.com","username":"ryan-tidwell"},"change_message_id":"2b7b4dee7783738ca1b08077730090c3ad8f22d3","unresolved":false,"context_lines":[{"line_number":69,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details.\"\"\""},{"line_number":70,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":71,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":72,"context_line":"        set_br_name(utils.gen_trunk_br_name(port[\u0027id\u0027]))"}],"source_content_type":"text/x-python","patch_set":22,"id":"9ad45d7e_d26dd633","line":72,"updated":"2016-08-08 20:43:40.000000000","message":"Just starting to play with some of Rossella\u0027s agent patches. Let\u0027s use the trunk_id instead of the port id here.","commit_id":"e46fc462367d1545a015ef2b3a569ce2af1c3175"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"ebc29cbe8138b4304438abf8330187623739c4ba","unresolved":false,"context_lines":[{"line_number":69,"context_line":"    \"\"\"If port is a trunk port, generate a bridge_name for its vif_details.\"\"\""},{"line_number":70,"context_line":"    port \u003d kwargs[\u0027port\u0027]"},{"line_number":71,"context_line":"    if \u0027trunk_details\u0027 in port:"},{"line_number":72,"context_line":"        set_br_name(utils.gen_trunk_br_name(port[\u0027id\u0027]))"}],"source_content_type":"text/x-python","patch_set":22,"id":"9ad45d7e_78e20905","line":72,"in_reply_to":"9ad45d7e_d26dd633","updated":"2016-08-08 21:23:13.000000000","message":"Done","commit_id":"e46fc462367d1545a015ef2b3a569ce2af1c3175"}],"neutron/services/trunk/plugin.py":[{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"0dfbd26d0b93ed93b519d79137767fe8587fb22e","unresolved":false,"context_lines":[{"line_number":32,"context_line":"from neutron.services.trunk import constants"},{"line_number":33,"context_line":"from neutron.services.trunk import exceptions as trunk_exc"},{"line_number":34,"context_line":"from neutron.services.trunk import rules"},{"line_number":35,"context_line":"from neutron.services.trunk.drivers import openvswitch  # noqa"},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"LOG \u003d logging.getLogger(__name__)"},{"line_number":38,"context_line":""}],"source_content_type":"text/x-python","patch_set":7,"id":"bacf61ea_c56330db","line":35,"updated":"2016-07-28 16:22:46.000000000","message":"pep8: non-alphabetical","commit_id":"e95a26c91b3191d4270ac53b42273686c9f89c49"}],"neutron/services/trunk/utils.py":[{"author":{"_account_id":4187,"name":"Ryan Tidwell","email":"rtidwell@suse.com","username":"ryan-tidwell"},"change_message_id":"814dc824af61264898530999ff64f130f3ac058e","unresolved":false,"context_lines":[{"line_number":15,"context_line":""},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"def gen_trunk_br_name(port_id):"},{"line_number":18,"context_line":"    return (\u0027tbr-\u0027 + port_id)[:constants.DEVICE_NAME_MAX_LEN - 1]"}],"source_content_type":"text/x-python","patch_set":13,"id":"bacf61ea_6da9b1df","line":18,"range":{"start_line":18,"start_character":13,"end_line":18,"end_character":16},"updated":"2016-08-02 22:52:02.000000000","message":"nit: maybe create a constant for the \"tbr\" prefix","commit_id":"9b43a8a6924bf3a70db4f0030295c53a88b26610"}],"neutron/tests/functional/services/trunk/test_plugin.py":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"7a2d7d552c551911fdb833078b6ce3281499c08c","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_10fe1838","line":22,"updated":"2016-08-09 15:31:07.000000000","message":"Any particular reason why you made this into a functional test? You register the agent for no clear reason and do system interaction with this test.","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"9bd77045d69a466f9e85edfbd1a143a68e709ae2","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_fbe7e5e1","line":22,"in_reply_to":"9ad45d7e_10fe1838","updated":"2016-08-09 15:42:23.000000000","message":"I register the agent so that the full portbinding process occurs after I set the port\u0027s HOST_ID. This functional test provides a more end-to-end test rather than just UTs.","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"4b1c559aa6fd14e860f1c68fba2c73190ea1146a","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_d4958be4","line":22,"in_reply_to":"9ad45d7e_3950605e","updated":"2016-08-09 18:06:39.000000000","message":"No, let\u0027s leave it. I am happy with your justification.","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"f3b297f3a04255dee907d1110a9a2f6c74275f9c","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_3950605e","line":22,"in_reply_to":"9ad45d7e_435a8751","updated":"2016-08-09 17:45:34.000000000","message":"At least in this test as it is, the portbinding doesn\u0027t occur unless the ovs agent has been registered, which makes me believe that as a UT without registering an ovs agent the full portbinding won\u0027t occur either.\n\nEssentially the UT coverage we have now:\n- assert that the OVS mech_driver properly emits a notification during portbinding to allow setting bridge_name\n- assert that the OVS trunk driver properly subscribes to the callback and passes a bridge_name to the trigger method\n\nThat coverage alone should surface most failures. One thing this test does add is making sure that the driver doesn\u0027t set bridge_name for non-trunks, but that could easily be mocked in a new UT.\n\nSo, should I just drop this functional test and add one more mocked-out UT for that?","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"f5427733c031c60e4b223cca0c1671c35cc31ddd","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_435a8751","line":22,"in_reply_to":"9ad45d7e_43d727bf","updated":"2016-08-09 17:04:49.000000000","message":"if not mistaken, this could just move under unit, remove the agent registration and it should just work as is.","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"},{"author":{"_account_id":4187,"name":"Ryan Tidwell","email":"rtidwell@suse.com","username":"ryan-tidwell"},"change_message_id":"e164fb68946cf1ea48b93f357b200a454fc4eb38","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_43d727bf","line":22,"in_reply_to":"9ad45d7e_56e1a899","updated":"2016-08-09 17:01:24.000000000","message":"This does seem like overkill. Although maybe not truly a unit test, this behavior could be asserted with less overhead when written as a unit test perhaps employing some mocks if necessary.","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"0798f45909ab14583598689feaf2d29ec75e2509","unresolved":false,"context_lines":[{"line_number":19,"context_line":"from neutron.tests.unit.plugins.ml2 import base as ml2_test_base"},{"line_number":20,"context_line":""},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"class TestTrunkServicePlugin(ml2_test_base.ML2TestFramework):"},{"line_number":23,"context_line":""},{"line_number":24,"context_line":"    def setUp(self):"},{"line_number":25,"context_line":"        super(TestTrunkServicePlugin, self).setUp()"}],"source_content_type":"text/x-python","patch_set":24,"id":"9ad45d7e_56e1a899","line":22,"in_reply_to":"9ad45d7e_fbe7e5e1","updated":"2016-08-09 15:57:48.000000000","message":"I question the value though, there\u0027s no failure that would surface in the test though is there?","commit_id":"f697cddda60f6d8d61ad147634ebc53f6d6139f7"}],"neutron/tests/unit/plugins/ml2/drivers/openvswitch/mech_driver/test_mech_openvswitch.py":[{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"31607067795a58dfc46b8f00d034d383e30d4b31","unresolved":false,"context_lines":[{"line_number":20,"context_line":"from neutron.callbacks import registry"},{"line_number":21,"context_line":"from neutron.extensions import portbindings"},{"line_number":22,"context_line":"from neutron.plugins.ml2 import driver_api as api"},{"line_number":23,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.agent.common \\"},{"line_number":24,"context_line":"    import constants as a_const"},{"line_number":25,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.mech_driver \\"},{"line_number":26,"context_line":"    import mech_openvswitch"}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_13a43db7","line":23,"updated":"2016-07-26 00:15:25.000000000","message":"ditto","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"6cbd74dcbcc78a20656f8ea9da3acb71cbbc5e15","unresolved":false,"context_lines":[{"line_number":20,"context_line":"from neutron.callbacks import registry"},{"line_number":21,"context_line":"from neutron.extensions import portbindings"},{"line_number":22,"context_line":"from neutron.plugins.ml2 import driver_api as api"},{"line_number":23,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.agent.common \\"},{"line_number":24,"context_line":"    import constants as a_const"},{"line_number":25,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.mech_driver \\"},{"line_number":26,"context_line":"    import mech_openvswitch"}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_5714b3e7","line":23,"in_reply_to":"dada55a8_13a43db7","updated":"2016-07-26 16:52:34.000000000","message":"Done","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"31607067795a58dfc46b8f00d034d383e30d4b31","unresolved":false,"context_lines":[{"line_number":22,"context_line":"from neutron.plugins.ml2 import driver_api as api"},{"line_number":23,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.agent.common \\"},{"line_number":24,"context_line":"    import constants as a_const"},{"line_number":25,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.mech_driver \\"},{"line_number":26,"context_line":"    import mech_openvswitch"},{"line_number":27,"context_line":"from neutron.tests.unit.plugins.ml2 import _test_mech_agent as base"},{"line_number":28,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_73abf989","line":25,"updated":"2016-07-26 00:15:25.000000000","message":"ditto","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"6cbd74dcbcc78a20656f8ea9da3acb71cbbc5e15","unresolved":false,"context_lines":[{"line_number":22,"context_line":"from neutron.plugins.ml2 import driver_api as api"},{"line_number":23,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.agent.common \\"},{"line_number":24,"context_line":"    import constants as a_const"},{"line_number":25,"context_line":"from neutron.plugins.ml2.drivers.openvswitch.mech_driver \\"},{"line_number":26,"context_line":"    import mech_openvswitch"},{"line_number":27,"context_line":"from neutron.tests.unit.plugins.ml2 import _test_mech_agent as base"},{"line_number":28,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"dada55a8_b7d36f37","line":25,"in_reply_to":"dada55a8_73abf989","updated":"2016-07-26 16:52:34.000000000","message":"Done","commit_id":"fbe0ad34203874ebdb5ac982273bb5f6d936f5d9"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"0363ff4194e1e8428a3538d02a3a96c3fe8fc421","unresolved":false,"context_lines":[{"line_number":17,"context_line":"from oslo_config import cfg"},{"line_number":18,"context_line":""},{"line_number":19,"context_line":"from neutron.callbacks import events"},{"line_number":20,"context_line":"from neutron.callbacks import exceptions as cb_exc"},{"line_number":21,"context_line":"from neutron.callbacks import registry"},{"line_number":22,"context_line":"from neutron.extensions import portbindings"},{"line_number":23,"context_line":"from neutron.plugins.ml2 import driver_api as api"}],"source_content_type":"text/x-python","patch_set":23,"id":"9ad45d7e_3b9526e4","line":20,"updated":"2016-08-09 00:33:05.000000000","message":"doh","commit_id":"8675becbf2697a2ee957fd6be757c3e99c021332"}],"neutron/tests/unit/services/trunk/drivers/openvswitch/test_driver.py":[{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"bee84caf84d75438cc7d1915fdcc041efab84bea","unresolved":false,"context_lines":[{"line_number":40,"context_line":"    def test_vif_details_bridge_name_handler_registration(self,"},{"line_number":41,"context_line":"                                                          mock_gen_br_name):"},{"line_number":42,"context_line":"        mock_gen_br_name.return_value \u003d \u0027fake-trunk-br-name\u0027"},{"line_number":43,"context_line":"        ovs_driver.register()"},{"line_number":44,"context_line":"        test_trigger \u003d mock.Mock()"},{"line_number":45,"context_line":"        registry.notify(agent_consts.OVS_BRIDGE_NAME, events.PRECOMMIT_CREATE,"},{"line_number":46,"context_line":"                        test_trigger, **{\u0027port\u0027: {\u0027trunk_details\u0027: \u0027foo\u0027,"}],"source_content_type":"text/x-python","patch_set":4,"id":"bacf61ea_ce30a5a1","line":43,"range":{"start_line":43,"start_character":8,"end_line":43,"end_character":29},"updated":"2016-07-27 20:19:16.000000000","message":"will remove this line due to it being moved into setUp","commit_id":"212b944110ab280c87a27c58f0908fc9209e0800"},{"author":{"_account_id":19999,"name":"Trinath Somanchi","email":"trinaths.dev@gmail.com","username":"trinath-somanchi"},"change_message_id":"b03e6580a6d6a486f6b8b51354d103a075aa6f89","unresolved":false,"context_lines":[{"line_number":49,"context_line":"    @mock.patch(\u0027neutron.services.trunk.drivers.openvswitch.driver.\u0027"},{"line_number":50,"context_line":"                \u0027add_subports_handler\u0027)"},{"line_number":51,"context_line":"    def test_subports_after_create_handler_registeration(self, handler_mock):"},{"line_number":52,"context_line":"        subport \u003d trunk_objects.SubPort(context\u003dself.context,"},{"line_number":53,"context_line":"                                        port_id\u003duuidutils.generate_uuid(),"},{"line_number":54,"context_line":"                                        segmentation_id\u003d2,"},{"line_number":55,"context_line":"                                        segmentation_type\u003d\u0027vlan\u0027)"}],"source_content_type":"text/x-python","patch_set":5,"id":"bacf61ea_0a272913","line":52,"updated":"2016-07-28 11:12:41.000000000","message":"typo: _registration ! as above","commit_id":"abdfb5582f5f3bda7b78b9cd62839c988443aa7a"},{"author":{"_account_id":16800,"name":"Rawlin Peters","email":"rawlin.peters@hpe.com","username":"rawlin"},"change_message_id":"3de2cebe1b769485e37680cfb36d7852f244eb06","unresolved":false,"context_lines":[{"line_number":49,"context_line":"    @mock.patch(\u0027neutron.services.trunk.drivers.openvswitch.driver.\u0027"},{"line_number":50,"context_line":"                \u0027add_subports_handler\u0027)"},{"line_number":51,"context_line":"    def test_subports_after_create_handler_registeration(self, handler_mock):"},{"line_number":52,"context_line":"        subport \u003d trunk_objects.SubPort(context\u003dself.context,"},{"line_number":53,"context_line":"                                        port_id\u003duuidutils.generate_uuid(),"},{"line_number":54,"context_line":"                                        segmentation_id\u003d2,"},{"line_number":55,"context_line":"                                        segmentation_type\u003d\u0027vlan\u0027)"}],"source_content_type":"text/x-python","patch_set":5,"id":"bacf61ea_4521009f","line":52,"in_reply_to":"bacf61ea_0a272913","updated":"2016-07-28 16:00:30.000000000","message":"That is code from the previously-depended-upon patch. I\u0027ve noted the typo in that patch.","commit_id":"abdfb5582f5f3bda7b78b9cd62839c988443aa7a"}]}
