rdkafka 0.7.0 → 0.8.0

data/lib/rdkafka/producer/delivery_report.rb

@@ -10,11 +10,16 @@ module Rdkafka
       # @return [Integer]
       attr_reader :offset
 
+      # Error in case happen during produce.
+      # @return [string]
+      attr_reader :error
+
       private
 
-      def initialize(partition, offset)
+      def initialize(partition, offset, error = nil)
         @partition = partition
         @offset = offset
+        @error = error
       end
     end
   end

data/lib/rdkafka/producer.rb

@@ -37,10 +37,25 @@ module Rdkafka
 
     # Close this producer and wait for the internal poll queue to empty.
     def close
+      return unless @native_kafka
+
       # Indicate to polling thread that we're closing
       @closing = true
       # Wait for the polling thread to finish up
       @polling_thread.join
+      Rdkafka::Bindings.rd_kafka_destroy(@native_kafka)
+      @native_kafka = nil
+    end
+
+    # Partition count for a given topic.
+    # NOTE: If 'allow.auto.create.topics' is set to true in the broker, the topic will be auto-created after returning nil.
+    #
+    # @param topic [String] The topic name.
+    #
+    # @return partition count [Integer,nil]
+    #
+    def partition_count(topic)
+      Rdkafka::Metadata.new(@native_kafka, topic).topics&.select { |x| x[:topic_name] == topic }&.dig(0, :partition_count)
     end
 
     # Produces a message to a Kafka topic. The message is added to rdkafka's queue, call {DeliveryHandle#wait wait} on the returned delivery handle to make sure it is delivered.
@@ -58,7 +73,7 @@ module Rdkafka
     # @raise [RdkafkaError] When adding the message to rdkafka's queue failed
     #
     # @return [DeliveryHandle] Delivery handle that can be used to wait for the result of producing this message
-    def produce(topic:, payload: nil, key: nil, partition: nil, timestamp: nil, headers: nil)
+    def produce(topic:, payload: nil, key: nil, partition: nil, partition_key: nil, timestamp: nil, headers: nil)
       # Start by checking and converting the input
 
       # Get payload length
@@ -75,9 +90,15 @@ module Rdkafka
                    key.bytesize
                  end
 
-      # If partition is nil use -1 to let Kafka set the partition based
-      # on the key/randomly if there is no key
-      partition = -1 if partition.nil?
+      if partition_key
+        partition_count = partition_count(topic)
+        # If the topic is not present, set to -1
+        partition = Rdkafka::Bindings.partitioner(partition_key, partition_count) if partition_count
+      end
+
+      # If partition is nil, use -1 to let librdafka set the partition randomly or
+      # based on the key when present.
+      partition ||= -1
 
       # If timestamp is nil use 0 and let Kafka set one. If an integer or time
       # use it.

data/lib/rdkafka/version.rb

@@ -1,5 +1,5 @@
 module Rdkafka
-  VERSION = "0.7.0"
-  LIBRDKAFKA_VERSION = "1.2.0"
-  LIBRDKAFKA_SOURCE_SHA256 = "eedde1c96104e4ac2d22a4230e34f35dd60d53976ae2563e3dd7c27190a96859"
+  VERSION = "0.8.0"
+  LIBRDKAFKA_VERSION = "1.4.0"
+  LIBRDKAFKA_SOURCE_SHA256 = "ae27ea3f3d0d32d29004e7f709efbba2666c5383a107cc45b3a1949486b2eb84"
 end

data/lib/rdkafka.rb

@@ -8,6 +8,7 @@ require "rdkafka/consumer/message"
 require "rdkafka/consumer/partition"
 require "rdkafka/consumer/topic_partition_list"
 require "rdkafka/error"
+require "rdkafka/metadata"
 require "rdkafka/producer"
 require "rdkafka/producer/delivery_handle"
 require "rdkafka/producer/delivery_report"

data/rdkafka.gemspec

@@ -19,7 +19,7 @@ Gem::Specification.new do |gem|
 
   gem.add_dependency 'ffi', '~> 1.9'
   gem.add_dependency 'mini_portile2', '~> 2.1'
-  gem.add_dependency 'rake', '~> 12.3'
+  gem.add_dependency 'rake', '>= 12.3'
 
   gem.add_development_dependency 'pry', '~> 0.10'
   gem.add_development_dependency 'rspec', '~> 3.5'

data/spec/rdkafka/bindings_spec.rb

@@ -1,4 +1,5 @@
 require "spec_helper"
+require 'zlib'
 
 describe Rdkafka::Bindings do
   it "should load librdkafka" do
@@ -7,12 +8,12 @@ describe Rdkafka::Bindings do
 
   describe ".lib_extension" do
     it "should know the lib extension for darwin" do
-      expect(Gem::Platform.local).to receive(:os).and_return("darwin-aaa")
+      stub_const('RbConfig::CONFIG', 'host_os' =>'darwin')
       expect(Rdkafka::Bindings.lib_extension).to eq "dylib"
     end
 
     it "should know the lib extension for linux" do
-      expect(Gem::Platform.local).to receive(:os).and_return("linux")
+      stub_const('RbConfig::CONFIG', 'host_os' =>'linux')
       expect(Rdkafka::Bindings.lib_extension).to eq "so"
     end
   end
@@ -60,6 +61,23 @@ describe Rdkafka::Bindings do
     end
   end
 
+  describe "partitioner" do
+    let(:partition_key) { ('a'..'z').to_a.shuffle.take(15).join('') }
+    let(:partition_count) { rand(50) + 1 }
+
+    it "should return the same partition for a similar string and the same partition count" do
+      result_1 = Rdkafka::Bindings.partitioner(partition_key, partition_count)
+      result_2 = Rdkafka::Bindings.partitioner(partition_key, partition_count)
+      expect(result_1).to eq(result_2)
+    end
+
+    it "should match the old partitioner" do
+      result_1 = Rdkafka::Bindings.partitioner(partition_key, partition_count)
+      result_2 = (Zlib.crc32(partition_key) % partition_count)
+      expect(result_1).to eq(result_2)
+    end
+  end
+
   describe "stats callback" do
     context "without a stats callback" do
       it "should do nothing" do

data/spec/rdkafka/config_spec.rb

@@ -50,7 +50,9 @@ describe Rdkafka::Config do
     end
 
     it "should create a consumer with valid config" do
-      expect(rdkafka_config.consumer).to be_a Rdkafka::Consumer
+      consumer = rdkafka_config.consumer
+      expect(consumer).to be_a Rdkafka::Consumer
+      consumer.close
     end
 
     it "should raise an error when creating a consumer with invalid config" do
@@ -76,7 +78,9 @@ describe Rdkafka::Config do
     end
 
     it "should create a producer with valid config" do
-      expect(rdkafka_config.producer).to be_a Rdkafka::Producer
+      producer = rdkafka_config.producer
+      expect(producer).to be_a Rdkafka::Producer
+      producer.close
     end
 
     it "should raise an error when creating a producer with invalid config" do

data/spec/rdkafka/consumer/message_spec.rb

@@ -1,7 +1,8 @@
 require "spec_helper"
 
 describe Rdkafka::Consumer::Message do
-  let(:native_topic) { new_native_topic }
+  let(:native_client) { new_native_client }
+  let(:native_topic) { new_native_topic(native_client: native_client) }
   let(:payload) { nil }
   let(:key) { nil }
   let(:native_message) do
@@ -24,6 +25,10 @@ describe Rdkafka::Consumer::Message do
     end
   end
 
+  after(:each) do
+    Rdkafka::Bindings.rd_kafka_destroy(native_client)
+  end
+
   subject { Rdkafka::Consumer::Message.new(native_message) }
 
   before do

data/spec/rdkafka/consumer_spec.rb

@@ -6,7 +6,10 @@ describe Rdkafka::Consumer do
   let(:consumer) { config.consumer }
   let(:producer) { config.producer }
 
-  describe "#subscripe, #unsubscribe and #subscription" do
+  after { consumer.close }
+  after { producer.close }
+
+  describe "#subscribe, #unsubscribe and #subscription" do
     it "should subscribe, unsubscribe and return the subscription" do
       expect(consumer.subscription).to be_empty
 
@@ -88,7 +91,6 @@ describe Rdkafka::Consumer do
         # 8. ensure that message is successfully consumed
         records = consumer.poll(timeout)
         expect(records).not_to be_nil
-        consumer.commit
       end
     end
 
@@ -205,8 +207,6 @@ describe Rdkafka::Consumer do
         expect(records&.payload).to eq "payload c"
         records = consumer.poll(timeout)
         expect(records).to be_nil
-
-        consumer.commit
       end
     end
   end
@@ -313,11 +313,11 @@ describe Rdkafka::Consumer do
       }.to raise_error TypeError
     end
 
-    context "with a commited consumer" do
+    context "with a committed consumer" do
       before :all do
-        # Make sure there are some message
-        producer = rdkafka_config.producer
+        # Make sure there are some messages.
         handles = []
+        producer = rdkafka_config.producer
         10.times do
           (0..2).each do |i|
             handles << producer.produce(
@@ -329,6 +329,7 @@ describe Rdkafka::Consumer do
           end
         end
         handles.each(&:wait)
+        producer.close
       end
 
       before do
@@ -389,20 +390,26 @@ describe Rdkafka::Consumer do
 
       describe "#store_offset" do
         before do
+          config = {}
           config[:'enable.auto.offset.store'] = false
           config[:'enable.auto.commit'] = false
-          consumer.subscribe("consume_test_topic")
-          wait_for_assignment(consumer)
+          @new_consumer = rdkafka_config(config).consumer
+          @new_consumer.subscribe("consume_test_topic")
+          wait_for_assignment(@new_consumer)
+        end
+
+        after do
+          @new_consumer.close
         end
 
         it "should store the offset for a message" do
-          consumer.store_offset(message)
-          consumer.commit
+          @new_consumer.store_offset(message)
+          @new_consumer.commit
 
           list = Rdkafka::Consumer::TopicPartitionList.new.tap do |list|
             list.add_topic("consume_test_topic", [0, 1, 2])
           end
-          partitions = consumer.committed(list).to_h["consume_test_topic"]
+          partitions = @new_consumer.committed(list).to_h["consume_test_topic"]
           expect(partitions).not_to be_nil
           expect(partitions[message.partition].offset).to eq(message.offset + 1)
         end
@@ -410,7 +417,7 @@ describe Rdkafka::Consumer do
         it "should raise an error with invalid input" do
           allow(message).to receive(:partition).and_return(9999)
           expect {
-            consumer.store_offset(message)
+            @new_consumer.store_offset(message)
           }.to raise_error Rdkafka::RdkafkaError
         end
       end
@@ -554,12 +561,12 @@ describe Rdkafka::Consumer do
         payload:   "payload 1",
         key:       "key 1"
       ).wait
-
       consumer.subscribe("consume_test_topic")
-      message = consumer.poll(5000)
-      expect(message).to be_a Rdkafka::Consumer::Message
+      message = consumer.each {|m| break m}
 
-      # Message content is tested in producer spec
+      expect(message).to be_a Rdkafka::Consumer::Message
+      expect(message.payload).to eq('payload 1')
+      expect(message.key).to eq('key 1')
     end
 
     it "should raise an error when polling fails" do

data/spec/rdkafka/error_spec.rb

@@ -71,15 +71,15 @@ describe Rdkafka::RdkafkaError do
     end
 
     it "should not equal another error with a different error code" do
-      expect(subject).to eq Rdkafka::RdkafkaError.new(20, "Error explanation")
+      expect(subject).not_to eq Rdkafka::RdkafkaError.new(20, "Error explanation")
     end
 
     it "should not equal another error with a different message" do
-      expect(subject).to eq Rdkafka::RdkafkaError.new(10, "Different error explanation")
+      expect(subject).not_to eq Rdkafka::RdkafkaError.new(10, "Different error explanation")
     end
 
     it "should not equal another error with no message" do
-      expect(subject).to eq Rdkafka::RdkafkaError.new(10)
+      expect(subject).not_to eq Rdkafka::RdkafkaError.new(10)
     end
   end
 end

data/spec/rdkafka/producer/delivery_report_spec.rb

@@ -1,7 +1,7 @@
 require "spec_helper"
 
 describe Rdkafka::Producer::DeliveryReport do
-  subject { Rdkafka::Producer::DeliveryReport.new(2, 100) }
+  subject { Rdkafka::Producer::DeliveryReport.new(2, 100, "error") }
 
   it "should get the partition" do
     expect(subject.partition).to eq 2
@@ -10,4 +10,8 @@ describe Rdkafka::Producer::DeliveryReport do
   it "should get the offset" do
     expect(subject.offset).to eq 100
   end
+
+  it "should get the error" do
+    expect(subject.error).to eq "error"
+  end
 end

data/spec/rdkafka/producer_spec.rb

@@ -2,10 +2,13 @@ require "spec_helper"
 
 describe Rdkafka::Producer do
   let(:producer) { rdkafka_config.producer }
+  let(:consumer) { rdkafka_config.consumer }
 
   after do
     # Registry should always end up being empty
     expect(Rdkafka::Producer::DeliveryHandle::REGISTRY).to be_empty
+    producer.close
+    consumer.close
   end
 
   context "delivery callback" do
@@ -27,6 +30,7 @@ describe Rdkafka::Producer do
     it "should call the callback when a message is delivered" do
       @callback_called = false
 
+
       producer.delivery_callback = lambda do |report|
         expect(report).not_to be_nil
         expect(report.partition).to eq 1
@@ -42,7 +46,10 @@ describe Rdkafka::Producer do
       )
 
       # Wait for it to be delivered
-      handle.wait(max_wait_timeout: 5)
+      handle.wait(max_wait_timeout: 15)
+
+      # Join the producer thread.
+      producer.close
 
       # Callback should have been called
       expect(@callback_called).to be true
@@ -55,7 +62,7 @@ describe Rdkafka::Producer do
         payload: "payload",
         key:     "key"
      )
-    }.to raise_error ArgumentError, "missing keyword: topic"
+    }.to raise_error ArgumentError, /missing keyword: [\:]?topic/
   end
 
   it "should produce a message" do
@@ -82,14 +89,15 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
     expect(message.partition).to eq 1
     expect(message.payload).to eq "payload"
     expect(message.key).to eq "key"
     # Since api.version.request is on by default we will get
     # the message creation timestamp if it's not set.
-    expect(message.timestamp).to be_within(5).of(Time.now)
+    expect(message.timestamp).to be_within(10).of(Time.now)
   end
 
   it "should produce a message with a specified partition" do
@@ -105,12 +113,47 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
     expect(message.partition).to eq 1
     expect(message.key).to eq "key partition"
   end
 
+  it "should produce a message to the same partition with a similar partition key" do
+    # Avoid partitioner collisions.
+    while true
+      key = ('a'..'z').to_a.shuffle.take(10).join('')
+      partition_key = ('a'..'z').to_a.shuffle.take(10).join('')
+      partition_count = producer.partition_count('partitioner_test_topic')
+      break if (Zlib.crc32(key) % partition_count) != (Zlib.crc32(partition_key) % partition_count)
+    end
+
+    # Produce a message with key, partition_key and key + partition_key
+    messages = [{key: key}, {partition_key: partition_key}, {key: key, partition_key: partition_key}]
+
+    messages = messages.map do |m|
+      handle = producer.produce(
+        topic:     "partitioner_test_topic",
+        payload:   "payload partition",
+        key:       m[:key],
+        partition_key: m[:partition_key]
+      )
+      report = handle.wait(max_wait_timeout: 5)
+
+      wait_for_message(
+        topic: "partitioner_test_topic",
+        delivery_report: report,
+      )
+    end
+
+    expect(messages[0].partition).not_to eq(messages[2].partition)
+    expect(messages[1].partition).to eq(messages[2].partition)
+    expect(messages[0].key).to eq key
+    expect(messages[1].key).to be_nil
+    expect(messages[2].key).to eq key
+  end
+
   it "should produce a message with utf-8 encoding" do
     handle = producer.produce(
       topic:   "produce_test_topic",
@@ -122,7 +165,8 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
 
     expect(message.partition).to eq 1
@@ -154,7 +198,8 @@ describe Rdkafka::Producer do
       # Consume message and verify it's content
       message = wait_for_message(
         topic: "produce_test_topic",
-        delivery_report: report
+        delivery_report: report,
+        consumer: consumer
       )
 
       expect(message.partition).to eq 2
@@ -174,7 +219,8 @@ describe Rdkafka::Producer do
       # Consume message and verify it's content
       message = wait_for_message(
         topic: "produce_test_topic",
-        delivery_report: report
+        delivery_report: report,
+        consumer: consumer
       )
 
       expect(message.partition).to eq 2
@@ -193,7 +239,8 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
 
     expect(message.key).to be_nil
@@ -210,7 +257,8 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
 
     expect(message.key).to eq "key no payload"
@@ -229,7 +277,8 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
 
     expect(message.payload).to eq "payload headers"
@@ -251,7 +300,8 @@ describe Rdkafka::Producer do
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
 
     expect(message.payload).to eq "payload headers"
@@ -284,11 +334,17 @@ describe Rdkafka::Producer do
     # Fork, produce a message, send the report over a pipe and
     # wait for and check the message in the main process.
 
+    # Kernel#fork is not available in JRuby
+    skip if defined?(JRUBY_VERSION)
+
     reader, writer = IO.pipe
 
     fork do
       reader.close
 
+      # Avoids sharing the socket between processes.
+      producer = rdkafka_config.producer
+
       handle = producer.produce(
         topic:   "produce_test_topic",
         payload: "payload-forked",
@@ -296,7 +352,6 @@ describe Rdkafka::Producer do
       )
 
       report = handle.wait(max_wait_timeout: 5)
-      producer.close
 
       report_json = JSON.generate(
         "partition" => report.partition,
@@ -304,20 +359,24 @@ describe Rdkafka::Producer do
       )
 
       writer.write(report_json)
+      writer.close
+      producer.close
     end
 
     writer.close
-
     report_hash = JSON.parse(reader.read)
     report = Rdkafka::Producer::DeliveryReport.new(
       report_hash["partition"],
       report_hash["offset"]
     )
 
+    reader.close
+
     # Consume message and verify it's content
     message = wait_for_message(
       topic: "produce_test_topic",
-      delivery_report: report
+      delivery_report: report,
+      consumer: consumer
     )
     expect(message.partition).to eq 0
     expect(message.payload).to eq "payload-forked"

data/spec/spec_helper.rb

@@ -7,6 +7,14 @@ require "pry"
 require "rspec"
 require "rdkafka"
 
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 3 --if-not-exists --topic consume_test_topic`
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 3 --if-not-exists --topic empty_test_topic`
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 3 --if-not-exists --topic load_test_topic`
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 3 --if-not-exists --topic produce_test_topic`
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 3 --if-not-exists --topic rake_test_topic`
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 3 --if-not-exists --topic watermarks_test_topic`
+`docker-compose exec kafka kafka-topics --create --zookeeper zookeeper:2181 --replication-factor 1 --partitions 25 --if-not-exists --topic partitioner_test_topic`
+
 def rdkafka_config(config_overrides={})
   config = {
     :"api.version.request" => false,
@@ -25,12 +33,12 @@ def rdkafka_config(config_overrides={})
   Rdkafka::Config.new(config)
 end
 
-def native_client
+def new_native_client
   config = rdkafka_config
   config.send(:native_kafka, config.send(:native_config), :rd_kafka_producer)
 end
 
-def new_native_topic(topic_name="topic_name")
+def new_native_topic(topic_name="topic_name", native_client: )
   Rdkafka::Bindings.rd_kafka_topic_new(
     native_client,
     topic_name,
@@ -39,7 +47,8 @@ def new_native_topic(topic_name="topic_name")
 end
 
 def wait_for_message(topic:, delivery_report:, timeout_in_seconds: 30, consumer: nil)
-  consumer = rdkafka_config.consumer if consumer.nil?
+  new_consumer = !!consumer
+  consumer ||= rdkafka_config.consumer
   consumer.subscribe(topic)
   timeout = Time.now.to_i + timeout_in_seconds
   loop do
@@ -53,6 +62,8 @@ def wait_for_message(topic:, delivery_report:, timeout_in_seconds: 30, consumer:
       return message
     end
   end
+ensure
+  consumer.close if new_consumer
 end
 
 def wait_for_assignment(consumer)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rdkafka
 version: !ruby/object:Gem::Version
-  version: 0.7.0
+  version: 0.8.0
 platform: ruby
 authors:
 - Thijs Cadier
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2019-09-21 00:00:00.000000000 Z
+date: 2020-06-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
@@ -42,14 +42,14 @@ dependencies:
   name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: '12.3'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: '12.3'
 - !ruby/object:Gem::Dependency
@@ -136,6 +136,7 @@ files:
 - lib/rdkafka/consumer/partition.rb
 - lib/rdkafka/consumer/topic_partition_list.rb
 - lib/rdkafka/error.rb
+- lib/rdkafka/metadata.rb
 - lib/rdkafka/producer.rb
 - lib/rdkafka/producer/delivery_handle.rb
 - lib/rdkafka/producer/delivery_report.rb
@@ -171,8 +172,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.6.2
+rubygems_version: 3.1.2
 signing_key: 
 specification_version: 4
 summary: The rdkafka gem is a modern Kafka client library for Ruby based on librdkafka.