)]}'
{"neutron/plugins/ml2/managers.py":[{"author":{"_account_id":7448,"name":"Carl Baldwin","email":"carl@ecbaldwin.net","username":"carl-baldwin"},"change_message_id":"2291b810655229696222fb1dd0fe85fec1a91389","unresolved":false,"context_lines":[{"line_number":385,"context_line":"        \"\"\""},{"line_number":386,"context_line":"        if context.current.get(\u0027vlan_transparent\u0027):"},{"line_number":387,"context_line":"            for driver in self.ordered_mech_drivers:"},{"line_number":388,"context_line":"                if driver.obj.check_vlan_transparency(context) is False:"},{"line_number":389,"context_line":"                    raise vlantransparent.VlanTransparencyDriverError()"},{"line_number":390,"context_line":""},{"line_number":391,"context_line":"    def _call_on_drivers(self, method_name, context,"}],"source_content_type":"text/x-python","patch_set":1,"id":"9ad45d7e_ebfcf273","line":388,"range":{"start_line":388,"start_character":62,"end_line":388,"end_character":71},"updated":"2016-08-10 17:23:31.000000000","message":"I\u0027m concerned. Wasn\u0027t the intend to default to False if a driver inherits the base implementation [1]? If it were not, I\u0027d think that method would\u0027ve been written to return True. Why is it okay to assume that any legacy driver is okay with vlan transparency?\n\n[1] https://github.com/openstack/neutron/blob/04b3b4dfa89bd265972193f0f7ba10fb96c0c4da/neutron/plugins/ml2/driver_api.py#L892","commit_id":"383e7d1f80c2555310659198faf966f8f361cd18"},{"author":{"_account_id":10652,"name":"Erik Colnick","email":"erikcolnick@gmail.com","username":"ecolnick"},"change_message_id":"404b1168b9f2975d2d47befa8ca8e656e1ae7836","unresolved":false,"context_lines":[{"line_number":385,"context_line":"        \"\"\""},{"line_number":386,"context_line":"        if context.current.get(\u0027vlan_transparent\u0027):"},{"line_number":387,"context_line":"            for driver in self.ordered_mech_drivers:"},{"line_number":388,"context_line":"                if driver.obj.check_vlan_transparency(context) is False:"},{"line_number":389,"context_line":"                    raise vlantransparent.VlanTransparencyDriverError()"},{"line_number":390,"context_line":""},{"line_number":391,"context_line":"    def _call_on_drivers(self, method_name, context,"}],"source_content_type":"text/x-python","patch_set":1,"id":"9ad45d7e_81e437e1","line":388,"range":{"start_line":388,"start_character":62,"end_line":388,"end_character":71},"in_reply_to":"9ad45d7e_e6c8a9aa","updated":"2016-08-10 18:08:15.000000000","message":"I am going to abandon this change.  I went back and re-read the spec and found that in the spec it is explicitly stated in the \u0027Work Items\u0027 that in \u0027the case that any driver does not have the attribute or has the attribute set to false, ML2 will decline to create VLAN transparent networks.\u0027[1]  That being the case, this approach is invalid.\n\n[1] \nhttp://specs.openstack.org/openstack/neutron-specs/specs/kilo/nfv-vlan-trunks.html#work-items","commit_id":"383e7d1f80c2555310659198faf966f8f361cd18"},{"author":{"_account_id":10652,"name":"Erik Colnick","email":"erikcolnick@gmail.com","username":"ecolnick"},"change_message_id":"e72ee73a266cffe25f6ec992e4b9cfd35075af26","unresolved":false,"context_lines":[{"line_number":385,"context_line":"        \"\"\""},{"line_number":386,"context_line":"        if context.current.get(\u0027vlan_transparent\u0027):"},{"line_number":387,"context_line":"            for driver in self.ordered_mech_drivers:"},{"line_number":388,"context_line":"                if driver.obj.check_vlan_transparency(context) is False:"},{"line_number":389,"context_line":"                    raise vlantransparent.VlanTransparencyDriverError()"},{"line_number":390,"context_line":""},{"line_number":391,"context_line":"    def _call_on_drivers(self, method_name, context,"}],"source_content_type":"text/x-python","patch_set":1,"id":"9ad45d7e_e6c8a9aa","line":388,"range":{"start_line":388,"start_character":62,"end_line":388,"end_character":71},"in_reply_to":"9ad45d7e_ebfcf273","updated":"2016-08-10 17:35:34.000000000","message":"I am not sure that the original intent was to default to False.  The base implementation does not return anything at all and is just a \u0027pass\u0027.  Also notice that in the original accepted implementation [1] the logic was the same as in this patch.  It was only with the subsequent move from core to extension [2] that this was changed.  Lastly, the original spec states \"Plugins that have not been adapted to understand the property will ignore this flag and create the network regardless\" [3].  Both the l2pop and mech_sriov plugins have not been adapted (neither implements check_vlan_transparency) but neither plugin has any impact on vlan transparency either.  \n\n[1] https://review.openstack.org/#/c/158420/18/neutron/plugins/ml2/managers.py\n[2]\nhttps://review.openstack.org/#/c/169569/\n[3]\nhttp://specs.openstack.org/openstack/neutron-specs/specs/kilo/nfv-vlan-trunks.html#request","commit_id":"383e7d1f80c2555310659198faf966f8f361cd18"},{"author":{"_account_id":7448,"name":"Carl Baldwin","email":"carl@ecbaldwin.net","username":"carl-baldwin"},"change_message_id":"890bd6491ef5c56b3fd4cf15a1cce77b9d058499","unresolved":false,"context_lines":[{"line_number":385,"context_line":"        \"\"\""},{"line_number":386,"context_line":"        if context.current.get(\u0027vlan_transparent\u0027):"},{"line_number":387,"context_line":"            for driver in self.ordered_mech_drivers:"},{"line_number":388,"context_line":"                if driver.obj.check_vlan_transparency(context) is False:"},{"line_number":389,"context_line":"                    raise vlantransparent.VlanTransparencyDriverError()"},{"line_number":390,"context_line":""},{"line_number":391,"context_line":"    def _call_on_drivers(self, method_name, context,"}],"source_content_type":"text/x-python","patch_set":1,"id":"9ad45d7e_2616414c","line":388,"range":{"start_line":388,"start_character":62,"end_line":388,"end_character":71},"in_reply_to":"9ad45d7e_ebfcf273","updated":"2016-08-10 17:29:12.000000000","message":"I reread the bug and I see now why you want to revert to this behavior. However, it still seems odd to me to require a driver to explicitly declare it is not compatible with vlan transparency given that before we had the concept, many of them were not.","commit_id":"383e7d1f80c2555310659198faf966f8f361cd18"}]}