)]}'
{"src/connmap.cc":[{"author":{"_account_id":1000338,"name":"Abhi Dangeti","email":"abhinav@couchbase.com","username":"abhinavdangeti","avatars":[{"url":"https://www.gravatar.com/avatar/9007099855b1bf7d96e9c1a10bc9542e.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d32","height":32},{"url":"https://www.gravatar.com/avatar/9007099855b1bf7d96e9c1a10bc9542e.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d56","height":56},{"url":"https://www.gravatar.com/avatar/9007099855b1bf7d96e9c1a10bc9542e.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d100","height":100},{"url":"https://www.gravatar.com/avatar/9007099855b1bf7d96e9c1a10bc9542e.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d120","height":120}]},"change_message_id":"e07e24d946c1d8711479d49e715a8da869f5db28","unresolved":false,"context_lines":[{"line_number":782,"context_line":"        }"},{"line_number":783,"context_line":"    }"},{"line_number":784,"context_line":"    rlh.unlock();"},{"line_number":785,"context_line":""},{"line_number":786,"context_line":"    LockHolder lh(connsLock);"},{"line_number":787,"context_line":"    all.clear();"},{"line_number":788,"context_line":"    map_.clear();"}],"source_content_type":"text/x-c","patch_set":2,"id":"ddd9853f_a254200a","line":785,"updated":"2015-10-07 18:23:25.000000000","message":"This could introduce a leak (because of a race scenario), where we do not release the reference of a connection that was added to \"all\" after we\u0027ve attempted to release all its contents.","commit_id":"b5a6fef112e899fab8427ccb251380ef1ee632cd"},{"author":{"_account_id":1000097,"name":"Chiyoung Seo","email":"chiyoung@northscale.com","username":"chiyoung","avatars":[{"url":"https://www.gravatar.com/avatar/ec517f4126dd292151b452ce4d1edf31.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d32","height":32},{"url":"https://www.gravatar.com/avatar/ec517f4126dd292151b452ce4d1edf31.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d56","height":56},{"url":"https://www.gravatar.com/avatar/ec517f4126dd292151b452ce4d1edf31.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d100","height":100},{"url":"https://www.gravatar.com/avatar/ec517f4126dd292151b452ce4d1edf31.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d120","height":120}]},"change_message_id":"5a249773f45dadff88dcfde6a589e514876cf560","unresolved":false,"context_lines":[{"line_number":782,"context_line":"        }"},{"line_number":783,"context_line":"    }"},{"line_number":784,"context_line":"    rlh.unlock();"},{"line_number":785,"context_line":""},{"line_number":786,"context_line":"    LockHolder lh(connsLock);"},{"line_number":787,"context_line":"    all.clear();"},{"line_number":788,"context_line":"    map_.clear();"}],"source_content_type":"text/x-c","patch_set":2,"id":"ddd9853f_a2e34002","line":785,"in_reply_to":"ddd9853f_a254200a","updated":"2015-10-07 20:21:38.000000000","message":"Well, the racing issue that you mentioned can even happen in the old implementation too if we accept a new connection during the bucket shutdown. I think we need to make sure that any new connection attempts to a given bucket should be denied once that bucket is currently being destroyed. If I remember correctly, we already check this in memcached. Otherwise, we should fix it.","commit_id":"b5a6fef112e899fab8427ccb251380ef1ee632cd"},{"author":{"_account_id":1000966,"name":"Dave Rigby","email":"daver@couchbase.com","username":"drigby","avatars":[{"url":"https://www.gravatar.com/avatar/514e75a8d75cc1fcdb22433d445ae8f1.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d32","height":32},{"url":"https://www.gravatar.com/avatar/514e75a8d75cc1fcdb22433d445ae8f1.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d56","height":56},{"url":"https://www.gravatar.com/avatar/514e75a8d75cc1fcdb22433d445ae8f1.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d100","height":100},{"url":"https://www.gravatar.com/avatar/514e75a8d75cc1fcdb22433d445ae8f1.jpg?d\u003didenticon\u0026r\u003dpg\u0026s\u003d120","height":120}]},"change_message_id":"afbade011aa5371053710b350e77335f84d7f8f8","unresolved":false,"context_lines":[{"line_number":782,"context_line":"        }"},{"line_number":783,"context_line":"    }"},{"line_number":784,"context_line":"    rlh.unlock();"},{"line_number":785,"context_line":""},{"line_number":786,"context_line":"    LockHolder lh(connsLock);"},{"line_number":787,"context_line":"    all.clear();"},{"line_number":788,"context_line":"    map_.clear();"}],"source_content_type":"text/x-c","patch_set":2,"id":"ddd9853f_421ef424","line":785,"in_reply_to":"ddd9853f_a2e34002","updated":"2015-10-07 20:43:54.000000000","message":"I\u0027ll double-check tomorrow. If it\u0027s not already handled in memcached, we can create a simple atomic\u003cbool\u003e shutdown_in_progress and guard connection creation on that.","commit_id":"b5a6fef112e899fab8427ccb251380ef1ee632cd"}]}