)]}'
{"neutron/agent/linux/dhcp.py":[{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"0afeae6ffe858841a01ccdc877d951732d5b51b3","unresolved":false,"context_lines":[{"line_number":1128,"context_line":""},{"line_number":1129,"context_line":"            # ignore weird IPv6 prefixes"},{"line_number":1130,"context_line":"            if (subnet.ip_version \u003d\u003d constants.IP_VERSION_6 and"},{"line_number":1131,"context_line":"                    netaddr.IPNetwork(subnet.cidr).prefixlen \u003c 64):"},{"line_number":1132,"context_line":"                continue"},{"line_number":1133,"context_line":""},{"line_number":1134,"context_line":"            if gateway !\u003d subnet.gateway_ip:"}],"source_content_type":"text/x-python","patch_set":6,"id":"1f013ff3_7185fc59","line":1131,"updated":"2017-05-17 10:39:55.000000000","message":"What is weird about prefixes \u003c /64 except that they may be even more wasteful than a /64?\n\nIMO what you really want to skip is \u003e\u003d /127 because there is no space to squeeze in a gateway there. Even /126 seems a bit weird.","commit_id":"27422250efa4a428c23195de351f85955c069923"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"b57d36ba9b9692621ac913c762bebb76fdc9ca28","unresolved":false,"context_lines":[{"line_number":1128,"context_line":""},{"line_number":1129,"context_line":"            # ignore weird IPv6 prefixes"},{"line_number":1130,"context_line":"            if (subnet.ip_version \u003d\u003d constants.IP_VERSION_6 and"},{"line_number":1131,"context_line":"                    netaddr.IPNetwork(subnet.cidr).prefixlen \u003c 64):"},{"line_number":1132,"context_line":"                continue"},{"line_number":1133,"context_line":""},{"line_number":1134,"context_line":"            if gateway !\u003d subnet.gateway_ip:"}],"source_content_type":"text/x-python","patch_set":6,"id":"ff0f0b1f_01692924","line":1131,"in_reply_to":"1f013ff3_7185fc59","updated":"2017-05-17 14:47:16.000000000","message":"Right, I was just taking out a big hammer to test things, and following RFC 5375 Section B.1 (Considerations for Subnet Prefixes Shorter than /64) - \"An allocation of a prefix shorter then 64 bits to a node or interface is considered bad practice\".  Even Section B.2 recommends against longer than /64.\n\nOut of the prefix lengths in the test, /0, /127 and /128 are not allowed by the API.  /1 isn\u0027t really usable either, but we still allow it and anything else the netaddr library doesn\u0027t complain about.  Since the test is bypassing the API we need to have some check here.\n\nShould we be stricter?  For example, if the mode is SLAAC should we enforce a /64 since otherwise it won\u0027t work?  I haven\u0027t looked if we do that.  Is a /32 large enough?  IANA doesn\u0027t even allocate large than a /16 from what I know (/20 is considered extra large), so maybe we don\u0027t allow larger?","commit_id":"27422250efa4a428c23195de351f85955c069923"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"300ff0b6ce67531ad3a1a5fa91a75adc64181f71","unresolved":false,"context_lines":[{"line_number":1128,"context_line":""},{"line_number":1129,"context_line":"            # ignore weird IPv6 prefixes"},{"line_number":1130,"context_line":"            if (subnet.ip_version \u003d\u003d constants.IP_VERSION_6 and"},{"line_number":1131,"context_line":"                    netaddr.IPNetwork(subnet.cidr).prefixlen \u003c 64):"},{"line_number":1132,"context_line":"                continue"},{"line_number":1133,"context_line":""},{"line_number":1134,"context_line":"            if gateway !\u003d subnet.gateway_ip:"}],"source_content_type":"text/x-python","patch_set":6,"id":"ff0f0b1f_042137b5","line":1131,"in_reply_to":"ff0f0b1f_01692924","updated":"2017-05-17 15:13:39.000000000","message":"The thing is that any hardcoded limit would be artificial, based on guessing what deployment scenario users may come up with. Technically I think there is no reason to prevent any of them except the ones you mention, even if I agree that the most useful size will be somewhere between /64 and /12x, for some x \u003c\u003d 5. But I don\u0027t think there is any value in having different behaviour for a /3, /32, /48 or /63 than we have for a /64, so my vote would be to drop that limit.\n\nAnd the fact that the API only disallows /0, /127 and /128 seems to undermine my claim that having functional tests for these network sizes is pretty useless.","commit_id":"27422250efa4a428c23195de351f85955c069923"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"744a9631a3bd5df37be2bf1d340750ddcc586344","unresolved":false,"context_lines":[{"line_number":1128,"context_line":""},{"line_number":1129,"context_line":"            # ignore weird IPv6 prefixes"},{"line_number":1130,"context_line":"            if (subnet.ip_version \u003d\u003d constants.IP_VERSION_6 and"},{"line_number":1131,"context_line":"                    netaddr.IPNetwork(subnet.cidr).prefixlen \u003c 64):"},{"line_number":1132,"context_line":"                continue"},{"line_number":1133,"context_line":""},{"line_number":1134,"context_line":"            if gateway !\u003d subnet.gateway_ip:"}],"source_content_type":"text/x-python","patch_set":6,"id":"ff0f0b1f_d701b3b1","line":1131,"in_reply_to":"ff0f0b1f_042137b5","updated":"2017-05-17 15:24:45.000000000","message":"Ok, let me send an update that mostly duplicates the checks done in the DB code, we\u0027ll see if the func test likes it.","commit_id":"27422250efa4a428c23195de351f85955c069923"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"9ad3c8b01cf5946d58a44dde726a88d6089bd5f3","unresolved":false,"context_lines":[{"line_number":1130,"context_line":"                # This is duplicating some of the API checks already done,"},{"line_number":1131,"context_line":"                # but some of the functional tests call directly"},{"line_number":1132,"context_line":"                prefixlen \u003d netaddr.IPNetwork(subnet.cidr).prefixlen"},{"line_number":1133,"context_line":"                if prefixlen \u003c\u003d 1 or prefixlen \u003e\u003d 127:"},{"line_number":1134,"context_line":"                    continue"},{"line_number":1135,"context_line":"                modes \u003d [constants.IPV6_SLAAC, constants.DHCPV6_STATELESS]"},{"line_number":1136,"context_line":"                addr_mode \u003d getattr(subnet, \u0027ipv6_address_mode\u0027, None)"}],"source_content_type":"text/x-python","patch_set":7,"id":"ff0f0b1f_f1f95c4d","line":1133,"updated":"2017-05-17 20:30:07.000000000","message":"nit to myself: _validate_subnet() checks prefixlen \u003e 126\n\nStill not sure about the minimum check, the IPAM code only checks it\u0027s !\u003d zero","commit_id":"d3ac435452600da2d4afa190dc0ba071396e2506"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"6a915ebe54b16af41221d913a319834be1dc06c7","unresolved":false,"context_lines":[{"line_number":1130,"context_line":"                # This is duplicating some of the API checks already done,"},{"line_number":1131,"context_line":"                # but some of the functional tests call directly"},{"line_number":1132,"context_line":"                prefixlen \u003d netaddr.IPNetwork(subnet.cidr).prefixlen"},{"line_number":1133,"context_line":"                if prefixlen \u003d\u003d 0 or prefixlen \u003e 126:"},{"line_number":1134,"context_line":"                    continue"},{"line_number":1135,"context_line":"                modes \u003d [constants.IPV6_SLAAC, constants.DHCPV6_STATELESS]"},{"line_number":1136,"context_line":"                addr_mode \u003d getattr(subnet, \u0027ipv6_address_mode\u0027, None)"}],"source_content_type":"text/x-python","patch_set":8,"id":"ff0f0b1f_9c7a3786","line":1133,"updated":"2017-05-17 20:35:22.000000000","message":"Let\u0027s see if the test explodes...","commit_id":"7ad7584ce113bff21999d6ffd155334bf3d05d2f"},{"author":{"_account_id":7787,"name":"Kevin Benton","email":"kevin@benton.pub","username":"blak111"},"change_message_id":"55b8d872c48b013384f1419dd9a132091d1bc91f","unresolved":false,"context_lines":[{"line_number":1130,"context_line":"                # This is duplicating some of the API checks already done,"},{"line_number":1131,"context_line":"                # but some of the functional tests call directly"},{"line_number":1132,"context_line":"                prefixlen \u003d netaddr.IPNetwork(subnet.cidr).prefixlen"},{"line_number":1133,"context_line":"                if prefixlen \u003d\u003d 0 or prefixlen \u003e 126:"},{"line_number":1134,"context_line":"                    continue"},{"line_number":1135,"context_line":"                modes \u003d [constants.IPV6_SLAAC, constants.DHCPV6_STATELESS]"},{"line_number":1136,"context_line":"                addr_mode \u003d getattr(subnet, \u0027ipv6_address_mode\u0027, None)"}],"source_content_type":"text/x-python","patch_set":8,"id":"ff0f0b1f_0eb70d3d","line":1133,"updated":"2017-05-26 01:48:27.000000000","message":"fwiw you can do a range check in python like so\n\nif not 0 \u003c prefixlen \u003c 127:","commit_id":"7ad7584ce113bff21999d6ffd155334bf3d05d2f"}]}
