)]}'
{"specs/in_progress/symlinks.rst":[{"author":{"_account_id":18820,"name":"Peter Chng","email":"pchng@ca.ibm.com","username":"pchng"},"change_message_id":"d178d8ae8aa6d864c2395371f673810e9a407c08","unresolved":false,"context_lines":[{"line_number":94,"context_line":""},{"line_number":95,"context_line":"Given the scenarios above at T1 (i.e. during the post) it is possible some object"},{"line_number":96,"context_line":"servers can see a symlink and others a regular object, thus it is not possible"},{"line_number":97,"context_line":"to fail the POST of a symlink. Instead, the following behaviour will be"},{"line_number":98,"context_line":"utilized, the object server will always apply the POST whether the object is a"},{"line_number":99,"context_line":"symlink or a regular object. Next, we will still return an error to the client"},{"line_number":100,"context_line":"if the object server believes it has seen a symlink. In scenario A) this would"}],"source_content_type":"text/x-rst","patch_set":1,"id":"da85f559_08820401","line":97,"updated":"2015-11-08 18:40:04.000000000","message":"This part *might* be clearer if broken up into a numbered list, e.g.:\n\nInstead, the following behaviour will be utilized:\n\n1. The object server will always apply the POST...\n2. An error will still be returned to the client if the object server believes...\n  - In scenario A) this would imply.. \n  - In scenario B) this would imply...\n\nThis is just a matter of personal preference, however.","commit_id":"dae8c22149699d80f24b38cf298a35feeb3440e9"},{"author":{"_account_id":9625,"name":"Thiago da Silva","email":"thiagodasilva@gmail.com","username":"thiago"},"change_message_id":"26cf9bde7c590d2f2b2b3bbec42a983b5f7c8d12","unresolved":false,"context_lines":[{"line_number":98,"context_line":"utilized, the object server will always apply the POST whether the object is a"},{"line_number":99,"context_line":"symlink or a regular object. Next, we will still return an error to the client"},{"line_number":100,"context_line":"if the object server believes it has seen a symlink. In scenario A) this would"},{"line_number":101,"context_line":"imply the POST at T1 may fail but the update will indeed be applied to the"},{"line_number":102,"context_line":"regular object, which is the correct behaviour. In scenario B) this would imply"},{"line_number":103,"context_line":"the POST at T1 may fail but the update will indeed be applied to the symlink,"},{"line_number":104,"context_line":"which while not ideal is not incorrect behaviour per say, and the error"}],"source_content_type":"text/x-rst","patch_set":1,"id":"da85f559_95c01f2c","line":101,"updated":"2015-11-09 20:09:07.000000000","message":"the POST on both scenarios happen at T2, right?","commit_id":"dae8c22149699d80f24b38cf298a35feeb3440e9"},{"author":{"_account_id":18820,"name":"Peter Chng","email":"pchng@ca.ibm.com","username":"pchng"},"change_message_id":"d178d8ae8aa6d864c2395371f673810e9a407c08","unresolved":false,"context_lines":[{"line_number":101,"context_line":"imply the POST at T1 may fail but the update will indeed be applied to the"},{"line_number":102,"context_line":"regular object, which is the correct behaviour. In scenario B) this would imply"},{"line_number":103,"context_line":"the POST at T1 may fail but the update will indeed be applied to the symlink,"},{"line_number":104,"context_line":"which while not ideal is not incorrect behaviour per say, and the error"},{"line_number":105,"context_line":"returned to the application should cause it to apply the POST to the reference"},{"line_number":106,"context_line":"object and given the initial point raised earlier this is indeed desirable."},{"line_number":107,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"da85f559_c895dc40","line":104,"updated":"2015-11-08 18:40:04.000000000","message":"(sp) Should be \"not incorrect behaviour *per se*, ...\"","commit_id":"dae8c22149699d80f24b38cf298a35feeb3440e9"},{"author":{"_account_id":9625,"name":"Thiago da Silva","email":"thiagodasilva@gmail.com","username":"thiago"},"change_message_id":"26cf9bde7c590d2f2b2b3bbec42a983b5f7c8d12","unresolved":false,"context_lines":[{"line_number":103,"context_line":"the POST at T1 may fail but the update will indeed be applied to the symlink,"},{"line_number":104,"context_line":"which while not ideal is not incorrect behaviour per say, and the error"},{"line_number":105,"context_line":"returned to the application should cause it to apply the POST to the reference"},{"line_number":106,"context_line":"object and given the initial point raised earlier this is indeed desirable."},{"line_number":107,"context_line":""},{"line_number":108,"context_line":"The aim is for Swift symlinks to operate analogously to Unix symbolic"},{"line_number":109,"context_line":"links (except where it does not make sense to do so)."}],"source_content_type":"text/x-rst","patch_set":1,"id":"da85f559_957d7fb3","line":106,"updated":"2015-11-09 20:09:07.000000000","message":"this is still a bit confusing. Returning an error when things actually succeeded sounds scary. I\u0027m sorry I don\u0027t have a better suggestion atm, but I\u0027m just afraid this might bite us later. \n\nCan you describe the behavior with the consistency engine in scenario A? does it need to know about symlinks?","commit_id":"dae8c22149699d80f24b38cf298a35feeb3440e9"}]}
