)]}'
{"keystoneauth1/access/service_catalog.py":[{"author":{"_account_id":8482,"name":"Colleen Murphy","email":"colleen@gazlene.net","username":"krinkle"},"change_message_id":"0ceddd4f8fd5dd579ea634bad289fdbaa610a86b","unresolved":false,"context_lines":[{"line_number":55,"context_line":"        \"\"\""},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"    @staticmethod"},{"line_number":58,"context_line":"    def normalize_interface(interface):"},{"line_number":59,"context_line":"        \"\"\"Handle differences in the way v2 and v3 catalogs specify endpoint."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"        Both v2 and v3 must be able to handle the endpoint style of the other."}],"source_content_type":"text/x-python","patch_set":1,"id":"1f1a1f67_6b131c68","line":58,"updated":"2017-07-12 14:04:08.000000000","message":"Didn\u0027t we have this discussion already - that we can\u0027t change this interface since it was already released? This wasn\u0027t introduced by mordred, I see it in 2.21.0.","commit_id":"2c7d745adfb65a002696bfc7dd0fea9001a0798d"},{"author":{"_account_id":14070,"name":"Eric Fried","email":"openstack@fried.cc","username":"efried"},"change_message_id":"3dbb1b0f29f5bb16bfaffd65bd091a5410951f54","unresolved":false,"context_lines":[{"line_number":55,"context_line":"        \"\"\""},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"    @staticmethod"},{"line_number":58,"context_line":"    def normalize_interface(interface):"},{"line_number":59,"context_line":"        \"\"\"Handle differences in the way v2 and v3 catalogs specify endpoint."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"        Both v2 and v3 must be able to handle the endpoint style of the other."}],"source_content_type":"text/x-python","patch_set":1,"id":"1f1a1f67_e64763bc","line":58,"in_reply_to":"1f1a1f67_46a977dd","updated":"2017-07-12 14:20:14.000000000","message":"Just so we\u0027re clear, in order to break an existing consumer, that consumer would need to be extending ServiceCatalog and then invoking normalize_interface with... something (?) passed in as the `self` argument, right?","commit_id":"2c7d745adfb65a002696bfc7dd0fea9001a0798d"},{"author":{"_account_id":10608,"name":"Matthew Edmonds","email":"edmondsw@us.ibm.com","username":"edmondsw"},"change_message_id":"207058f90f00a4fdc6d9a42eb3a6f9bdba5d3bda","unresolved":false,"context_lines":[{"line_number":55,"context_line":"        \"\"\""},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"    @staticmethod"},{"line_number":58,"context_line":"    def normalize_interface(interface):"},{"line_number":59,"context_line":"        \"\"\"Handle differences in the way v2 and v3 catalogs specify endpoint."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"        Both v2 and v3 must be able to handle the endpoint style of the other."}],"source_content_type":"text/x-python","patch_set":1,"id":"1f1a1f67_e676433f","line":58,"in_reply_to":"1f1a1f67_46a977dd","updated":"2017-07-12 14:14:36.000000000","message":"Yes this is an interface change, but surely nobody was using this as-is. And even on the *extremely* slim change that someone was somehow using this, they must have noticed the bug and should have opened such, so shame on them. This feels like a case where the rule about interface changes hurts users more than it helps them, and should therefore have an exception.","commit_id":"2c7d745adfb65a002696bfc7dd0fea9001a0798d"},{"author":{"_account_id":2,"name":"Monty Taylor","email":"mordred@inaugust.com","username":"mordred"},"change_message_id":"908a24fe379b2720b5bc0948c3d0133c0b610e41","unresolved":false,"context_lines":[{"line_number":55,"context_line":"        \"\"\""},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"    @staticmethod"},{"line_number":58,"context_line":"    def normalize_interface(interface):"},{"line_number":59,"context_line":"        \"\"\"Handle differences in the way v2 and v3 catalogs specify endpoint."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"        Both v2 and v3 must be able to handle the endpoint style of the other."}],"source_content_type":"text/x-python","patch_set":1,"id":"1f1a1f67_46a977dd","line":58,"in_reply_to":"1f1a1f67_6b131c68","updated":"2017-07-12 14:08:09.000000000","message":"Yes. good catch. even though it\u0027s a bug, in this case it\u0027s an interface change, so we can\u0027t do it.","commit_id":"2c7d745adfb65a002696bfc7dd0fea9001a0798d"},{"author":{"_account_id":14070,"name":"Eric Fried","email":"openstack@fried.cc","username":"efried"},"change_message_id":"803ce05b23fe9f05a79832375a2a478bd69f0003","unresolved":false,"context_lines":[{"line_number":55,"context_line":"        \"\"\""},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"    @staticmethod"},{"line_number":58,"context_line":"    def normalize_interface(interface):"},{"line_number":59,"context_line":"        \"\"\"Handle differences in the way v2 and v3 catalogs specify endpoint."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"        Both v2 and v3 must be able to handle the endpoint style of the other."}],"source_content_type":"text/x-python","patch_set":1,"id":"1f1a1f67_f1d1face","line":58,"in_reply_to":"1f1a1f67_8a0864d9","updated":"2017-07-14 21:42:59.000000000","message":"Removed.  I may get around to proposing the \"remove `self` from staticmethod\" patch at some point.  Or not.  Seems like we\u0027ve got bigger fish to fry, and there\u0027s lots of resistance to fixing it.","commit_id":"2c7d745adfb65a002696bfc7dd0fea9001a0798d"},{"author":{"_account_id":8482,"name":"Colleen Murphy","email":"colleen@gazlene.net","username":"krinkle"},"change_message_id":"3d987f4b0fc78b081dfd0cfd0f987f8d8cb17b2d","unresolved":false,"context_lines":[{"line_number":55,"context_line":"        \"\"\""},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"    @staticmethod"},{"line_number":58,"context_line":"    def normalize_interface(interface):"},{"line_number":59,"context_line":"        \"\"\"Handle differences in the way v2 and v3 catalogs specify endpoint."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"        Both v2 and v3 must be able to handle the endpoint style of the other."}],"source_content_type":"text/x-python","patch_set":1,"id":"1f1a1f67_8a0864d9","line":58,"in_reply_to":"1f1a1f67_e64763bc","updated":"2017-07-14 09:56:15.000000000","message":"Here\u0027s where we got to when this came up a couple of weeks ago: http://eavesdrop.openstack.org/irclogs/%23openstack-keystone/%23openstack-keystone.2017-06-08.log.html#t2017-06-08T23:02:42\n\nI would like for this to be proposed in its own patch, as Morgan suggested in that conversation. That way a revert, if needed, would be minimally disruptive to the git history.\n\nIt was also suggested that we fix it by making it not a staticmethod - I\u0027m not sure about that proposal, since all the correctly-overriding methods in this module are also staticmethods so I think changing that would be overzealous. We can have that discussion on the separate patch and have Morgan and Jamie weigh in.\n\nI agree that no one could be reasonably using it the way it is. I also think that, from a certain point of view, all bugfixes are backwards incompatible if someone is depending on broken behavior - but bugs still need to be fixed.","commit_id":"2c7d745adfb65a002696bfc7dd0fea9001a0798d"}]}