replicator http usage logger crashes when canceling transient (_replicate) jobs

[nickva]

http post $DB/_replicate id=87dc15f42308e145f24d56b2dcb2f143+continuous+create_target cancel:='true'
HTTP/1.1 500 Internal Server Error
0
{
    "error": "badrpc",
    "reason": "{'EXIT',{{badrecord,httpdb},\n         [{couch_replicator_utils,url_from_type,2,\n                                  [{file,\"src/couch_replicator_utils.erl\"},\n                                   {line,308}]},\n          {couch_replicator_utils,check_endpoint_protocols,2,\n                                  [{file,\"src/couch_replicator_utils.erl\"},\n                                   {line,295}]},\n          {couch_replicator_utils,valid_endpoint_protocols_log,1,\n                                  [{file,\"src/couch_replicator_utils.erl\"},\n                                   {line,288}]},\n          {couch_replicator,replicate,2,\n                            [{file,\"src/couch_replicator.erl\"},{line,61}]}]}}"
}

[nickva]

Example of how it fails:

5:51:14.961838 <0.7661.146> couch_replicator_utils:url_from_type({rep,undefined,undefined,undefined,
     [{cancel,true},
      {checkpoint_interval,30000},
      {connection_timeout,30000},
      {http_connections,20},
      {ibrowse_options,[]},
      {id,{"4252c9ffc7584d919a6d3899088beb5a","+continuous+create_target"}},
      {retries,5},
      {socket_options,[{keepalive,true},{nodelay,false}]},
      {use_bulk_get,false},
      {use_checkpoints,true},
      {worker_batch_size,10},
      {worker_processes,4}],
     {user_ctx,<<"testy026-user1">>,
               [<<"_admin">>,<<"_reader">>,<<"_writer">>],
               <<"default">>},
     db,nil,undefined,null,
     {1694,706674,961761},
     []}, source)
15:51:14.963080 <0.7661.146> couch_replicator_utils:url_from_type/2 error {badrecord,
                                                                           httpdb}