grpc 1.30.2-universal-darwin

data/src/ruby/lib/grpc/generic/bidi_call.rb

@@ -0,0 +1,233 @@
+# Copyright 2015 gRPC authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+require 'forwardable'
+require_relative '../grpc'
+
+# GRPC contains the General RPC module.
+module GRPC
+  # The BiDiCall class orchestrates execution of a BiDi stream on a client or
+  # server.
+  class BidiCall
+    include Core::CallOps
+    include Core::StatusCodes
+    include Core::TimeConsts
+
+    # Creates a BidiCall.
+    #
+    # BidiCall should only be created after a call is accepted.  That means
+    # different things on a client and a server.  On the client, the call is
+    # accepted after call.invoke. On the server, this is after call.accept.
+    #
+    # #initialize cannot determine if the call is accepted or not; so if a
+    # call that's not accepted is used here, the error won't be visible until
+    # the BidiCall#run is called.
+    #
+    # deadline is the absolute deadline for the call.
+    #
+    # @param call [Call] the call used by the ActiveCall
+    # @param marshal [Function] f(obj)->string that marshal requests
+    # @param unmarshal [Function] f(string)->obj that unmarshals responses
+    # @param metadata_received [true|false] indicates if metadata has already
+    #     been received. Should always be true for server calls
+    def initialize(call, marshal, unmarshal, metadata_received: false,
+                   req_view: nil)
+      fail(ArgumentError, 'not a call') unless call.is_a? Core::Call
+      @call = call
+      @marshal = marshal
+      @op_notifier = nil  # signals completion on clients
+      @unmarshal = unmarshal
+      @metadata_received = metadata_received
+      @reads_complete = false
+      @writes_complete = false
+      @complete = false
+      @done_mutex = Mutex.new
+      @req_view = req_view
+    end
+
+    # Begins orchestration of the Bidi stream for a client sending requests.
+    #
+    # The method either returns an Enumerator of the responses, or accepts a
+    # block that can be invoked with each response.
+    #
+    # @param requests the Enumerable of requests to send
+    # @param set_input_stream_done [Proc] called back when we're done
+    #   reading the input stream
+    # @param set_output_stream_done [Proc] called back when we're done
+    #   sending data on the output stream
+    # @return an Enumerator of requests to yield
+    def run_on_client(requests,
+                      set_input_stream_done,
+                      set_output_stream_done,
+                      &blk)
+      @enq_th = Thread.new do
+        write_loop(requests, set_output_stream_done: set_output_stream_done)
+      end
+      read_loop(set_input_stream_done, &blk)
+    end
+
+    # Begins orchestration of the Bidi stream for a server generating replies.
+    #
+    # N.B. gen_each_reply is a func(Enumerable<Requests>)
+    #
+    # It takes an enumerable of requests as an arg, in case there is a
+    # relationship between the stream of requests and the stream of replies.
+    #
+    # This does not mean that must necessarily be one.  E.g, the replies
+    # produced by gen_each_reply could ignore the received_msgs
+    #
+    # @param [Proc] gen_each_reply generates the BiDi stream replies.
+    # @param [Enumerable] requests The enumerable of requests to run
+    def run_on_server(gen_each_reply, requests)
+      replies = nil
+
+      # Pass in the optional call object parameter if possible
+      if gen_each_reply.arity == 1
+        replies = gen_each_reply.call(requests)
+      elsif gen_each_reply.arity == 2
+        replies = gen_each_reply.call(requests, @req_view)
+      else
+        fail 'Illegal arity of reply generator'
+      end
+
+      write_loop(replies, is_client: false)
+    end
+
+    ##
+    # Read the next stream iteration
+    #
+    # @param [Proc] finalize_stream callback to call when the reads have been
+    #   completely read through.
+    # @param [Boolean] is_client If this is a client or server request
+    #
+    def read_next_loop(finalize_stream, is_client = false)
+      read_loop(finalize_stream, is_client: is_client)
+    end
+
+    private
+
+    END_OF_READS = :end_of_reads
+    END_OF_WRITES = :end_of_writes
+
+    # performs a read using @call.run_batch, ensures metadata is set up
+    def read_using_run_batch
+      ops = { RECV_MESSAGE => nil }
+      ops[RECV_INITIAL_METADATA] = nil unless @metadata_received
+      begin
+        batch_result = @call.run_batch(ops)
+        unless @metadata_received
+          @call.metadata = batch_result.metadata
+          @metadata_received = true
+        end
+        batch_result
+      rescue GRPC::Core::CallError => e
+        GRPC.logger.warn('bidi call: read_using_run_batch failed')
+        GRPC.logger.warn(e)
+        nil
+      end
+    end
+
+    # set_output_stream_done is relevant on client-side
+    def write_loop(requests, is_client: true, set_output_stream_done: nil)
+      GRPC.logger.debug('bidi-write-loop: starting')
+      count = 0
+      requests.each do |req|
+        GRPC.logger.debug("bidi-write-loop: #{count}")
+        count += 1
+        payload = @marshal.call(req)
+        # Fails if status already received
+        begin
+          @req_view.send_initial_metadata unless @req_view.nil?
+          @call.run_batch(SEND_MESSAGE => payload)
+        rescue GRPC::Core::CallError => e
+          # This is almost definitely caused by a status arriving while still
+          # writing. Don't re-throw the error
+          GRPC.logger.warn('bidi-write-loop: ended with error')
+          GRPC.logger.warn(e)
+          break
+        end
+      end
+      GRPC.logger.debug("bidi-write-loop: #{count} writes done")
+      if is_client
+        GRPC.logger.debug("bidi-write-loop: client sent #{count}, waiting")
+        begin
+          @call.run_batch(SEND_CLOSE_FROM_CLIENT => nil)
+        rescue GRPC::Core::CallError => e
+          GRPC.logger.warn('bidi-write-loop: send close failed')
+          GRPC.logger.warn(e)
+        end
+        GRPC.logger.debug('bidi-write-loop: done')
+      end
+      GRPC.logger.debug('bidi-write-loop: finished')
+    rescue StandardError => e
+      GRPC.logger.warn('bidi-write-loop: failed')
+      GRPC.logger.warn(e)
+      if is_client
+        @call.cancel_with_status(GRPC::Core::StatusCodes::UNKNOWN,
+                                 "GRPC bidi call error: #{e.inspect}")
+      else
+        raise e
+      end
+    ensure
+      set_output_stream_done.call if is_client
+    end
+
+    # Provides an enumerator that yields results of remote reads
+    def read_loop(set_input_stream_done, is_client: true)
+      return enum_for(:read_loop,
+                      set_input_stream_done,
+                      is_client: is_client) unless block_given?
+      GRPC.logger.debug('bidi-read-loop: starting')
+      begin
+        count = 0
+        # queue the initial read before beginning the loop
+        loop do
+          GRPC.logger.debug("bidi-read-loop: #{count}")
+          count += 1
+          batch_result = read_using_run_batch
+
+          # handle the next message
+          if batch_result.nil? || batch_result.message.nil?
+            GRPC.logger.debug("bidi-read-loop: null batch #{batch_result}")
+
+            if is_client
+              batch_result = @call.run_batch(RECV_STATUS_ON_CLIENT => nil)
+              @call.status = batch_result.status
+              @call.trailing_metadata = @call.status.metadata if @call.status
+              GRPC.logger.debug("bidi-read-loop: done status #{@call.status}")
+              batch_result.check_status
+            end
+
+            GRPC.logger.debug('bidi-read-loop: done reading!')
+            break
+          end
+
+          res = @unmarshal.call(batch_result.message)
+          yield res
+        end
+      rescue StandardError => e
+        GRPC.logger.warn('bidi: read-loop failed')
+        GRPC.logger.warn(e)
+        raise e
+      ensure
+        set_input_stream_done.call
+      end
+      GRPC.logger.debug('bidi-read-loop: finished')
+      # Make sure that the write loop is done before finishing the call.
+      # Note that blocking is ok at this point because we've already received
+      # a status
+      @enq_th.join if is_client
+    end
+  end
+end

data/src/ruby/lib/grpc/generic/client_stub.rb

@@ -0,0 +1,501 @@
+# Copyright 2015 gRPC authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+require_relative 'active_call'
+require_relative '../version'
+
+# GRPC contains the General RPC module.
+module GRPC
+  # rubocop:disable Metrics/ParameterLists
+
+  # ClientStub represents a client connection to a gRPC server, and can be used
+  # to send requests.
+  class ClientStub
+    include Core::StatusCodes
+    include Core::TimeConsts
+
+    # Default timeout is infinity.
+    DEFAULT_TIMEOUT = INFINITE_FUTURE
+
+    # setup_channel is used by #initialize to constuct a channel from its
+    # arguments.
+    def self.setup_channel(alt_chan, host, creds, channel_args = {})
+      unless alt_chan.nil?
+        fail(TypeError, '!Channel') unless alt_chan.is_a?(Core::Channel)
+        return alt_chan
+      end
+      if channel_args['grpc.primary_user_agent'].nil?
+        channel_args['grpc.primary_user_agent'] = ''
+      else
+        channel_args['grpc.primary_user_agent'] += ' '
+      end
+      channel_args['grpc.primary_user_agent'] += "grpc-ruby/#{VERSION}"
+      unless creds.is_a?(Core::ChannelCredentials) || creds.is_a?(Symbol)
+        fail(TypeError, '!ChannelCredentials or Symbol')
+      end
+      Core::Channel.new(host, channel_args, creds)
+    end
+
+    # Allows users of the stub to modify the propagate mask.
+    #
+    # This is an advanced feature for use when making calls to another gRPC
+    # server whilst running in the handler of an existing one.
+    attr_writer :propagate_mask
+
+    # Creates a new ClientStub.
+    #
+    # Minimally, a stub is created with the just the host of the gRPC service
+    # it wishes to access, e.g.,
+    #
+    #   my_stub = ClientStub.new(example.host.com:50505,
+    #                            :this_channel_is_insecure)
+    #
+    # If a channel_override argument is passed, it will be used as the
+    # underlying channel. Otherwise, the channel_args argument will be used
+    # to construct a new underlying channel.
+    #
+    # There are some specific keyword args that are not used to configure the
+    # channel:
+    #
+    # - :channel_override
+    # when present, this must be a pre-created GRPC::Core::Channel.  If it's
+    # present the host and arbitrary keyword arg areignored, and the RPC
+    # connection uses this channel.
+    #
+    # - :timeout
+    # when present, this is the default timeout used for calls
+    #
+    # @param host [String] the host the stub connects to
+    # @param creds [Core::ChannelCredentials|Symbol] the channel credentials, or
+    #     :this_channel_is_insecure, which explicitly indicates that the client
+    #     should be created with an insecure connection. Note: this argument is
+    #     ignored if the channel_override argument is provided.
+    # @param channel_override [Core::Channel] a pre-created channel
+    # @param timeout [Number] the default timeout to use in requests
+    # @param propagate_mask [Number] A bitwise combination of flags in
+    #     GRPC::Core::PropagateMasks. Indicates how data should be propagated
+    #     from parent server calls to child client calls if this client is being
+    #     used within a gRPC server.
+    # @param channel_args [Hash] the channel arguments. Note: this argument is
+    #     ignored if the channel_override argument is provided.
+    # @param interceptors [Array<GRPC::ClientInterceptor>] An array of
+    #     GRPC::ClientInterceptor objects that will be used for
+    #     intercepting calls before they are executed
+    #     Interceptors are an EXPERIMENTAL API.
+    def initialize(host, creds,
+                   channel_override: nil,
+                   timeout: nil,
+                   propagate_mask: nil,
+                   channel_args: {},
+                   interceptors: [])
+      @ch = ClientStub.setup_channel(channel_override, host, creds,
+                                     channel_args)
+      alt_host = channel_args[Core::Channel::SSL_TARGET]
+      @host = alt_host.nil? ? host : alt_host
+      @propagate_mask = propagate_mask
+      @timeout = timeout.nil? ? DEFAULT_TIMEOUT : timeout
+      @interceptors = InterceptorRegistry.new(interceptors)
+    end
+
+    # request_response sends a request to a GRPC server, and returns the
+    # response.
+    #
+    # == Flow Control ==
+    # This is a blocking call.
+    #
+    # * it does not return until a response is received.
+    #
+    # * the requests is sent only when GRPC core's flow control allows it to
+    #   be sent.
+    #
+    # == Errors ==
+    # An RuntimeError is raised if
+    #
+    # * the server responds with a non-OK status
+    #
+    # * the deadline is exceeded
+    #
+    # == Return Value ==
+    #
+    # If return_op is false, the call returns the response
+    #
+    # If return_op is true, the call returns an Operation, calling execute
+    # on the Operation returns the response.
+    #
+    # @param method [String] the RPC method to call on the GRPC server
+    # @param req [Object] the request sent to the server
+    # @param marshal [Function] f(obj)->string that marshals requests
+    # @param unmarshal [Function] f(string)->obj that unmarshals responses
+    # @param deadline [Time] (optional) the time the request should complete
+    # @param return_op [true|false] return an Operation if true
+    # @param parent [Core::Call] a prior call whose reserved metadata
+    #   will be propagated by this one.
+    # @param credentials [Core::CallCredentials] credentials to use when making
+    #   the call
+    # @param metadata [Hash] metadata to be sent to the server
+    # @return [Object] the response received from the server
+    def request_response(method, req, marshal, unmarshal,
+                         deadline: nil,
+                         return_op: false,
+                         parent: nil,
+                         credentials: nil,
+                         metadata: {})
+      c = new_active_call(method, marshal, unmarshal,
+                          deadline: deadline,
+                          parent: parent,
+                          credentials: credentials)
+      interception_context = @interceptors.build_context
+      intercept_args = {
+        method: method,
+        request: req,
+        call: c.interceptable,
+        metadata: metadata
+      }
+      if return_op
+        # return the operation view of the active_call; define #execute as a
+        # new method for this instance that invokes #request_response.
+        c.merge_metadata_to_send(metadata)
+        op = c.operation
+        op.define_singleton_method(:execute) do
+          interception_context.intercept!(:request_response, intercept_args) do
+            c.request_response(req, metadata: metadata)
+          end
+        end
+        op
+      else
+        interception_context.intercept!(:request_response, intercept_args) do
+          c.request_response(req, metadata: metadata)
+        end
+      end
+    end
+
+    # client_streamer sends a stream of requests to a GRPC server, and
+    # returns a single response.
+    #
+    # requests provides an 'iterable' of Requests. I.e. it follows Ruby's
+    # #each enumeration protocol. In the simplest case, requests will be an
+    # array of marshallable objects; in typical case it will be an Enumerable
+    # that allows dynamic construction of the marshallable objects.
+    #
+    # == Flow Control ==
+    # This is a blocking call.
+    #
+    # * it does not return until a response is received.
+    #
+    # * each requests is sent only when GRPC core's flow control allows it to
+    #   be sent.
+    #
+    # == Errors ==
+    # An RuntimeError is raised if
+    #
+    # * the server responds with a non-OK status
+    #
+    # * the deadline is exceeded
+    #
+    # == Return Value ==
+    #
+    # If return_op is false, the call consumes the requests and returns
+    # the response.
+    #
+    # If return_op is true, the call returns the response.
+    #
+    # @param method [String] the RPC method to call on the GRPC server
+    # @param requests [Object] an Enumerable of requests to send
+    # @param marshal [Function] f(obj)->string that marshals requests
+    # @param unmarshal [Function] f(string)->obj that unmarshals responses
+    # @param deadline [Time] (optional) the time the request should complete
+    # @param return_op [true|false] return an Operation if true
+    # @param parent [Core::Call] a prior call whose reserved metadata
+    #   will be propagated by this one.
+    # @param credentials [Core::CallCredentials] credentials to use when making
+    #   the call
+    # @param metadata [Hash] metadata to be sent to the server
+    # @return [Object|Operation] the response received from the server
+    def client_streamer(method, requests, marshal, unmarshal,
+                        deadline: nil,
+                        return_op: false,
+                        parent: nil,
+                        credentials: nil,
+                        metadata: {})
+      c = new_active_call(method, marshal, unmarshal,
+                          deadline: deadline,
+                          parent: parent,
+                          credentials: credentials)
+      interception_context = @interceptors.build_context
+      intercept_args = {
+        method: method,
+        requests: requests,
+        call: c.interceptable,
+        metadata: metadata
+      }
+      if return_op
+        # return the operation view of the active_call; define #execute as a
+        # new method for this instance that invokes #client_streamer.
+        c.merge_metadata_to_send(metadata)
+        op = c.operation
+        op.define_singleton_method(:execute) do
+          interception_context.intercept!(:client_streamer, intercept_args) do
+            c.client_streamer(requests)
+          end
+        end
+        op
+      else
+        interception_context.intercept!(:client_streamer, intercept_args) do
+          c.client_streamer(requests, metadata: metadata)
+        end
+      end
+    end
+
+    # server_streamer sends one request to the GRPC server, which yields a
+    # stream of responses.
+    #
+    # responses provides an enumerator over the streamed responses, i.e. it
+    # follows Ruby's #each iteration protocol.  The enumerator blocks while
+    # waiting for each response, stops when the server signals that no
+    # further responses will be supplied.  If the implicit block is provided,
+    # it is executed with each response as the argument and no result is
+    # returned.
+    #
+    # == Flow Control ==
+    # This is a blocking call.
+    #
+    # * the request is sent only when GRPC core's flow control allows it to
+    #   be sent.
+    #
+    # * the request will not complete until the server sends the final
+    #   response followed by a status message.
+    #
+    # == Errors ==
+    # An RuntimeError is raised if
+    #
+    # * the server responds with a non-OK status when any response is
+    # * retrieved
+    #
+    # * the deadline is exceeded
+    #
+    # == Return Value ==
+    #
+    # if the return_op is false, the return value is an Enumerator of the
+    # results, unless a block is provided, in which case the block is
+    # executed with each response.
+    #
+    # if return_op is true, the function returns an Operation whose #execute
+    # method runs server streamer call. Again, Operation#execute either
+    # calls the given block with each response or returns an Enumerator of the
+    # responses.
+    #
+    # == Keyword Args ==
+    #
+    # Unspecified keyword arguments are treated as metadata to be sent to the
+    # server.
+    #
+    # @param method [String] the RPC method to call on the GRPC server
+    # @param req [Object] the request sent to the server
+    # @param marshal [Function] f(obj)->string that marshals requests
+    # @param unmarshal [Function] f(string)->obj that unmarshals responses
+    # @param deadline [Time] (optional) the time the request should complete
+    # @param return_op [true|false]return an Operation if true
+    # @param parent [Core::Call] a prior call whose reserved metadata
+    #   will be propagated by this one.
+    # @param credentials [Core::CallCredentials] credentials to use when making
+    #   the call
+    # @param metadata [Hash] metadata to be sent to the server
+    # @param blk [Block] when provided, is executed for each response
+    # @return [Enumerator|Operation|nil] as discussed above
+    def server_streamer(method, req, marshal, unmarshal,
+                        deadline: nil,
+                        return_op: false,
+                        parent: nil,
+                        credentials: nil,
+                        metadata: {},
+                        &blk)
+      c = new_active_call(method, marshal, unmarshal,
+                          deadline: deadline,
+                          parent: parent,
+                          credentials: credentials)
+      interception_context = @interceptors.build_context
+      intercept_args = {
+        method: method,
+        request: req,
+        call: c.interceptable,
+        metadata: metadata
+      }
+      if return_op
+        # return the operation view of the active_call; define #execute
+        # as a new method for this instance that invokes #server_streamer
+        c.merge_metadata_to_send(metadata)
+        op = c.operation
+        op.define_singleton_method(:execute) do
+          interception_context.intercept!(:server_streamer, intercept_args) do
+            c.server_streamer(req, &blk)
+          end
+        end
+        op
+      else
+        interception_context.intercept!(:server_streamer, intercept_args) do
+          c.server_streamer(req, metadata: metadata, &blk)
+        end
+      end
+    end
+
+    # bidi_streamer sends a stream of requests to the GRPC server, and yields
+    # a stream of responses.
+    #
+    # This method takes an Enumerable of requests, and returns and enumerable
+    # of responses.
+    #
+    # == requests ==
+    #
+    # requests provides an 'iterable' of Requests. I.e. it follows Ruby's
+    # #each enumeration protocol. In the simplest case, requests will be an
+    # array of marshallable objects; in typical case it will be an
+    # Enumerable that allows dynamic construction of the marshallable
+    # objects.
+    #
+    # == responses ==
+    #
+    # This is an enumerator of responses.  I.e, its #next method blocks
+    # waiting for the next response.  Also, if at any point the block needs
+    # to consume all the remaining responses, this can be done using #each or
+    # #collect.  Calling #each or #collect should only be done if
+    # the_call#writes_done has been called, otherwise the block will loop
+    # forever.
+    #
+    # == Flow Control ==
+    # This is a blocking call.
+    #
+    # * the call completes when the next call to provided block returns
+    #   false
+    #
+    # * the execution block parameters are two objects for sending and
+    #   receiving responses, each of which blocks waiting for flow control.
+    #   E.g, calles to bidi_call#remote_send will wait until flow control
+    #   allows another write before returning; and obviously calls to
+    #   responses#next block until the next response is available.
+    #
+    # == Termination ==
+    #
+    # As well as sending and receiving messages, the block passed to the
+    # function is also responsible for:
+    #
+    # * calling bidi_call#writes_done to indicate no further reqs will be
+    #   sent.
+    #
+    # * returning false if once the bidi stream is functionally completed.
+    #
+    # Note that response#next will indicate that there are no further
+    # responses by throwing StopIteration, but can only happen either
+    # if bidi_call#writes_done is called.
+    #
+    # To properly terminate the RPC, the responses should be completely iterated
+    # through; one way to do this is to loop on responses#next until no further
+    # responses are available.
+    #
+    # == Errors ==
+    # An RuntimeError is raised if
+    #
+    # * the server responds with a non-OK status when any response is
+    # * retrieved
+    #
+    # * the deadline is exceeded
+    #
+    #
+    # == Return Value ==
+    #
+    # if the return_op is false, the return value is an Enumerator of the
+    # results, unless a block is provided, in which case the block is
+    # executed with each response.
+    #
+    # if return_op is true, the function returns an Operation whose #execute
+    # method runs the Bidi call. Again, Operation#execute either calls a
+    # given block with each response or returns an Enumerator of the
+    # responses.
+    #
+    # @param method [String] the RPC method to call on the GRPC server
+    # @param requests [Object] an Enumerable of requests to send
+    # @param marshal [Function] f(obj)->string that marshals requests
+    # @param unmarshal [Function] f(string)->obj that unmarshals responses
+    # @param deadline [Time] (optional) the time the request should complete
+    # @param return_op [true|false] return an Operation if true
+    # @param parent [Core::Call] a prior call whose reserved metadata
+    #   will be propagated by this one.
+    # @param credentials [Core::CallCredentials] credentials to use when making
+    #   the call
+    # @param metadata [Hash] metadata to be sent to the server
+    # @param blk [Block] when provided, is executed for each response
+    # @return [Enumerator|nil|Operation] as discussed above
+    def bidi_streamer(method, requests, marshal, unmarshal,
+                      deadline: nil,
+                      return_op: false,
+                      parent: nil,
+                      credentials: nil,
+                      metadata: {},
+                      &blk)
+      c = new_active_call(method, marshal, unmarshal,
+                          deadline: deadline,
+                          parent: parent,
+                          credentials: credentials)
+      interception_context = @interceptors.build_context
+      intercept_args = {
+        method: method,
+        requests: requests,
+        call: c.interceptable,
+        metadata: metadata
+      }
+      if return_op
+        # return the operation view of the active_call; define #execute
+        # as a new method for this instance that invokes #bidi_streamer
+        c.merge_metadata_to_send(metadata)
+        op = c.operation
+        op.define_singleton_method(:execute) do
+          interception_context.intercept!(:bidi_streamer, intercept_args) do
+            c.bidi_streamer(requests, &blk)
+          end
+        end
+        op
+      else
+        interception_context.intercept!(:bidi_streamer, intercept_args) do
+          c.bidi_streamer(requests, metadata: metadata, &blk)
+        end
+      end
+    end
+
+    private
+
+    # Creates a new active stub
+    #
+    # @param method [string] the method being called.
+    # @param marshal [Function] f(obj)->string that marshals requests
+    # @param unmarshal [Function] f(string)->obj that unmarshals responses
+    # @param parent [Grpc::Call] a parent call, available when calls are
+    #   made from server
+    # @param credentials [Core::CallCredentials] credentials to use when making
+    #   the call
+    def new_active_call(method, marshal, unmarshal,
+                        deadline: nil,
+                        parent: nil,
+                        credentials: nil)
+      deadline = from_relative_time(@timeout) if deadline.nil?
+      # Provide each new client call with its own completion queue
+      call = @ch.create_call(parent, # parent call
+                             @propagate_mask, # propagation options
+                             method,
+                             nil, # host use nil,
+                             deadline)
+      call.set_credentials! credentials unless credentials.nil?
+      ActiveCall.new(call, marshal, unmarshal, deadline,
+                     started: false)
+    end
+  end
+end