rdkafka 0.16.0 → 0.17.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 07f1bd822870a6f28cd441b95870683442b3d5857f12d85e04057fa29a29121a
-  data.tar.gz: f734cec92e64dfdfbb6a7e3e1875ad402982aeea6d9e8e14befeb50144f2c125
+  metadata.gz: af536d53de701980f7ca5f5f426c3ea2450b4a1e46568c84edef9f5223544206
+  data.tar.gz: d7bfb39e6992112b84254b07cccced37730613c06ff58e9831990403c05c012c
 SHA512:
-  metadata.gz: 644f1435d988fd962cae3031d5d7c7a31de2d618e80def8f48913bda6309e08b81ff62741933b9ada5c40f59f79502e0a9dfee09efcc57fe7845269f178d1738
-  data.tar.gz: 3f83530ade65c38064b1c3a9922c173f973f57b49544fce3fd645a1c6a47ee18506450b33299e6417510f052123114fced56689a6ce2c59abd7d4be11b752c48
+  metadata.gz: 8801ec759340b9643a3d4971f21c0ab0563fdb7a416d2efeffa2e77de8525ebb69c92dcb394ceaae94513f0fd120acf41ff3f4a0efd405eae1438d8bb9835ee2
+  data.tar.gz: 12c937def332a678270d914645e2a0af6c06ffabdfc54284974cf386a68647411482df792e2289fea29473b904ef242309e574704426d355ce57a578aabd25a0

data/.ruby-version

@@ -1 +1 @@
-3.3.3
+3.3.4

data/CHANGELOG.md

@@ -1,5 +1,15 @@
 # Rdkafka Changelog
 
+## 0.17.0 (2024-08-03)
+- [Feature] Add `#seek_by` to be able to seek for a message by topic, partition and offset (zinahia)
+- [Enhancement] Update `librdkafka` to `2.4.0`
+- [Enhancement] Support ability to release patches to librdkafka.
+- [Change] Remove old producer timeout API warnings.
+- [Fix] Switch to local release of librdkafka to mitigate its unavailability.
+
+## 0.16.1 (2024-07-10)
+- [Fix] Switch to local release of librdkafka to mitigate its unavailability.
+
 ## 0.16.0 (2024-06-13)
 - **[Breaking]** Retire support for Ruby 2.7.
 - **[Breaking]** Messages without headers returned by `#poll` contain frozen empty hash.
@@ -18,6 +28,9 @@
 - [Fix] Background logger stops working after forking causing memory leaks (mensfeld)
 - [Fix] Fix bogus case/when syntax. Levels 1, 2, and 6 previously defaulted to UNKNOWN (jjowdy)
 
+## 0.15.2 (2024-07-10)
+- [Fix] Switch to local release of librdkafka to mitigate its unavailability.
+
 ## 0.15.1 (2024-01-30)
 - [Enhancement] Provide support for Nix OS (alexandriainfantino)
 - [Enhancement] Replace `rd_kafka_offset_store` with `rd_kafka_offsets_store` (mensfeld)
@@ -43,6 +56,9 @@
 - [Enhancement] Increase the `#lag` and `#query_watermark_offsets` default timeouts from 100ms to 1000ms. This will compensate for network glitches and remote clusters operations (mensfeld)
 - [Change] Use `SecureRandom.uuid` instead of `random` for test consumer groups (mensfeld)
 
+## 0.14.1 (2024-07-10)
+- [Fix] Switch to local release of librdkafka to mitigate its unavailability.
+
 ## 0.14.0 (2023-11-21)
 - [Enhancement] Add `raise_response_error` flag to the `Rdkafka::AbstractHandle`.
 - [Enhancement] Allow for setting `statistics_callback` as nil to reset predefined settings configured by a different gem (mensfeld)
@@ -60,6 +76,9 @@
 - [Change] Update Kafka Docker with Confluent KRaft (mensfeld)
 - [Change] Update librdkafka repo reference from edenhill to confluentinc (mensfeld)
 
+## 0.13.1 (2024-07-10)
+- [Fix] Switch to local release of librdkafka to mitigate its unavailability.
+
 ## 0.13.0 (2023-07-24)
 - Support cooperative sticky partition assignment in the rebalance callback (methodmissing)
 - Support both string and symbol header keys (ColinDKelley)
@@ -76,6 +95,9 @@
 - Make metadata request timeout configurable (mensfeld)
 - call_on_partitions_assigned and call_on_partitions_revoked only get a tpl passed in (thijsc)
 
+## 0.12.1 (2024-07-11)
+- [Fix] Switch to local release of librdkafka to mitigate its unavailability.
+
 ## 0.12.0 (2022-06-17)
 - Bumps librdkafka to 1.9.0
 - Fix crash on empty partition key (mensfeld)

data/README.md

@@ -161,12 +161,13 @@ bundle exec rake produce_messages
 
 ## Versions
 
-| rdkafka-ruby | librdkafka |
-|-|-|
-| 0.16.0 (2024-06-13) | 2.3.0 (2023-10-25) |
-| 0.15.0 (2023-12-03) | 2.3.0 (2023-10-25) |
-| 0.14.0 (2023-11-21) | 2.2.0 (2023-07-12) |
-| 0.13.0 (2023-07-24) | 2.0.2 (2023-01-20) |
-| 0.12.0 (2022-06-17) | 1.9.0 (2022-06-16) |
-| 0.11.0 (2021-11-17) | 1.8.2 (2021-10-18) |
-| 0.10.0 (2021-09-07) | 1.5.0 (2020-07-20) |
+| rdkafka-ruby | librdkafka | patches |
+|-|-|-|
+| 0.17.0 (2024-08-03) | 2.4.0 (2024-05-07) | no |
+| 0.16.0 (2024-06-13) | 2.3.0 (2023-10-25) | no |
+| 0.15.0 (2023-12-03) | 2.3.0 (2023-10-25) | no |
+| 0.14.0 (2023-11-21) | 2.2.0 (2023-07-12) | no |
+| 0.13.0 (2023-07-24) | 2.0.2 (2023-01-20) | no |
+| 0.12.0 (2022-06-17) | 1.9.0 (2022-06-16) | no |
+| 0.11.0 (2021-11-17) | 1.8.2 (2021-10-18) | no |
+| 0.10.0 (2021-09-07) | 1.5.0 (2020-07-20) | no |

data/docker-compose.yml

@@ -3,7 +3,7 @@ version: '2'
 services:
   kafka:
     container_name: kafka
-    image: confluentinc/cp-kafka:7.6.1
+    image: confluentinc/cp-kafka:7.7.0
 
     ports:
       - 9092:9092

data/ext/README.md

@@ -1,7 +1,7 @@
 # Ext
 
-This gem depends on the `librdkafka` C library. It is downloaded when
-this gem is installed.
+This gem depends on the `librdkafka` C library. It is downloaded, stored in
+`dist/` directory, and checked into source control.
 
 To update the `librdkafka` version follow the following steps:
 
@@ -9,8 +9,9 @@ To update the `librdkafka` version follow the following steps:
   version number and asset checksum for `tar.gz`.
 * Change the version in `lib/rdkafka/version.rb`
 * Change the `sha256` in `lib/rdkafka/version.rb`
-* Run `bundle exec rake` in the `ext` directory to download and build
-  the new version
+* Run `bundle exec rake dist:download` in the `ext` directory to download the
+  new release and place it in the `dist/` for you
+* Run `bundle exec rake` in the `ext` directory to build the new version
 * Run `docker-compose pull` in the main gem directory to ensure the docker
   images used by the tests and run `docker-compose up`
 * Finally, run `bundle exec rspec` in the main gem directory to execute

data/ext/Rakefile

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 
 require File.expand_path('../../lib/rdkafka/version', __FILE__)
+require "digest"
 require "fileutils"
 require "open-uri"
 
@@ -22,12 +23,16 @@ task :default => :clean do
       ENV["LDFLAGS"] = "-L#{homebrew_prefix}/lib" unless ENV["LDFLAGS"]
     end
 
+    releases = File.expand_path(File.join(File.dirname(__FILE__), '../dist'))
+
     recipe.files << {
-      :url => "https://codeload.github.com/edenhill/librdkafka/tar.gz/v#{Rdkafka::LIBRDKAFKA_VERSION}",
+      :url => "file://#{releases}/librdkafka_#{Rdkafka::LIBRDKAFKA_VERSION}.tar.gz",
       :sha256 => Rdkafka::LIBRDKAFKA_SOURCE_SHA256
     }
     recipe.configure_options = ["--host=#{recipe.host}"]
 
+    recipe.patch_files = Dir[File.join(releases, 'patches', "*.patch")].sort
+
     # Disable using libc regex engine in favor of the embedded one
     # The default regex engine of librdkafka does not always work exactly as most of the users
     # would expect, hence this flag allows for changing it to the other one
@@ -69,6 +74,42 @@ task :clean do
   FileUtils.rm_rf File.join(File.dirname(__FILE__), "tmp")
 end
 
+namespace :dist do
+  task :dir do
+    ENV["RDKAFKA_DIST_PATH"] ||= File.expand_path(File.join(File.dirname(__FILE__), '..', 'dist'))
+  end
+
+  task :file => "dist:dir" do
+    ENV["RDKAFKA_DIST_FILE"] ||= File.join(ENV["RDKAFKA_DIST_PATH"], "librdkafka_#{Rdkafka::LIBRDKAFKA_VERSION}.tar.gz")
+  end
+
+  task :clean => "dist:file" do
+    Dir.glob(File.join("#{ENV['RDKAFKA_DIST_PATH']}", "*")).each do |filename|
+      next if filename.include? ENV["RDKAFKA_DIST_FILE"]
+
+      FileUtils.rm_rf filename
+    end
+  end
+
+  task :download => "dist:file" do
+    version = Rdkafka::LIBRDKAFKA_VERSION
+    librdkafka_download = "https://codeload.github.com/confluentinc/librdkafka/tar.gz/v#{version}"
+
+    URI.open(librdkafka_download) do |file|
+      filename = ENV["RDKAFKA_DIST_FILE"]
+      data     = file.read
+
+      if Digest::SHA256.hexdigest(data) != Rdkafka::LIBRDKAFKA_SOURCE_SHA256
+        raise "SHA256 does not match downloaded file"
+      end
+
+      File.write(filename, data)
+    end
+  end
+
+  task :update => %w[dist:download dist:clean]
+end
+
 namespace :build do
   desc "Build librdkafka at the given git sha or tag"
   task :git, [:ref] do |task, args|
@@ -76,8 +117,9 @@ namespace :build do
     version = "git-#{ref}"
 
     recipe = MiniPortile.new("librdkafka", version)
-    recipe.files << "https://github.com/edenhill/librdkafka/archive/#{ref}.tar.gz"
+    recipe.files << "https://github.com/confluentinc/librdkafka/archive/#{ref}.tar.gz"
     recipe.configure_options = ["--host=#{recipe.host}","--enable-static", "--enable-zstd"]
+    recipe.patch_files = Dir[File.join(releases, 'patches', "*.patch")].sort
     recipe.cook
 
     ext = recipe.host.include?("darwin") ? "dylib" : "so"

data/lib/rdkafka/abstract_handle.rb

@@ -16,9 +16,6 @@ module Rdkafka
     REGISTRY = {}
     # Default wait timeout is 31 years
     MAX_WAIT_TIMEOUT_FOREVER = 10_000_000_000
-    # Deprecation message for wait_timeout argument in wait method
-    WAIT_TIMEOUT_DEPRECATION_MESSAGE = "The 'wait_timeout' argument is deprecated and will be removed in future versions without replacement. " \
-      "We don't rely on it's value anymore. Please refactor your code to remove references to it."
 
     private_constant :MAX_WAIT_TIMEOUT_FOREVER
 
@@ -59,16 +56,13 @@ module Rdkafka
     #
     # @param max_wait_timeout [Numeric, nil] Amount of time to wait before timing out.
     #   If this is nil we will wait forever
-    # @param wait_timeout [nil] deprecated
     # @param raise_response_error [Boolean] should we raise error when waiting finishes
     #
     # @return [Object] Operation-specific result
     #
     # @raise [RdkafkaError] When the operation failed
     # @raise [WaitTimeoutError] When the timeout has been reached and the handle is still pending
-    def wait(max_wait_timeout: 60, wait_timeout: nil, raise_response_error: true)
-      Kernel.warn(WAIT_TIMEOUT_DEPRECATION_MESSAGE) unless wait_timeout.nil?
-
+    def wait(max_wait_timeout: 60, raise_response_error: true)
       timeout = max_wait_timeout ? monotonic_now + max_wait_timeout : MAX_WAIT_TIMEOUT_FOREVER
 
       @mutex.synchronize do

data/lib/rdkafka/consumer.rb

@@ -423,6 +423,19 @@ module Rdkafka
     # @return [nil]
     # @raise [RdkafkaError] When seeking fails
     def seek(message)
+      seek_by(message.topic, message.partition, message.offset)
+    end
+
+    # Seek to a particular message by providing the topic, partition and offset.
+    # The next poll on the topic/partition will return the
+    # message at the given offset.
+    #
+    # @param topic [String] The topic in which to seek
+    # @param partition [Integer] The partition number to seek
+    # @param offset [Integer] The partition offset to seek
+    # @return [nil]
+    # @raise [RdkafkaError] When seeking fails
+    def seek_by(topic, partition, offset)
       closed_consumer_check(__method__)
 
       # rd_kafka_offset_store is one of the few calls that does not support
@@ -430,14 +443,14 @@ module Rdkafka
       native_topic = @native_kafka.with_inner do |inner|
         Rdkafka::Bindings.rd_kafka_topic_new(
           inner,
-          message.topic,
+          topic,
           nil
         )
       end
       response = Rdkafka::Bindings.rd_kafka_seek(
         native_topic,
-        message.partition,
-        message.offset,
+        partition,
+        offset,
         0 # timeout
       )
       if response != 0

data/lib/rdkafka/version.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 module Rdkafka
-  VERSION = "0.16.0"
-  LIBRDKAFKA_VERSION = "2.3.0"
-  LIBRDKAFKA_SOURCE_SHA256 = "2d49c35c77eeb3d42fa61c43757fcbb6a206daa560247154e60642bcdcc14d12"
+  VERSION = "0.17.0"
+  LIBRDKAFKA_VERSION = "2.4.0"
+  LIBRDKAFKA_SOURCE_SHA256 = "d645e47d961db47f1ead29652606a502bdd2a880c85c1e060e94eea040f1a19a"
 end

data/spec/rdkafka/abstract_handle_spec.rb

@@ -80,7 +80,6 @@ describe Rdkafka::AbstractHandle do
       let(:pending_handle) { true }
 
       it "should wait until the timeout and then raise an error" do
-        expect(Kernel).not_to receive(:warn)
         expect {
           subject.wait(max_wait_timeout: 0.1)
         }.to raise_error Rdkafka::AbstractHandle::WaitTimeoutError, /test_operation/
@@ -90,22 +89,15 @@ describe Rdkafka::AbstractHandle do
     context 'when pending_handle false' do
       let(:pending_handle) { false }
 
-      it 'should show a deprecation warning when wait_timeout is set' do
-        expect(Kernel).to receive(:warn).with(Rdkafka::AbstractHandle::WAIT_TIMEOUT_DEPRECATION_MESSAGE)
-        subject.wait(wait_timeout: 0.1)
-      end
-
       context "without error" do
         let(:result) { 1 }
 
         it "should return a result" do
-          expect(Kernel).not_to receive(:warn)
           wait_result = subject.wait
           expect(wait_result).to eq(result)
         end
 
         it "should wait without a timeout" do
-          expect(Kernel).not_to receive(:warn)
           wait_result = subject.wait(max_wait_timeout: nil)
           expect(wait_result).to eq(result)
         end
@@ -115,7 +107,6 @@ describe Rdkafka::AbstractHandle do
         let(:response) { 20 }
 
         it "should raise an rdkafka error" do
-          expect(Kernel).not_to receive(:warn)
           expect {
             subject.wait
           }.to raise_error Rdkafka::RdkafkaError

data/spec/rdkafka/admin_spec.rb

@@ -34,7 +34,7 @@ describe Rdkafka::Admin do
   describe '#describe_errors' do
     let(:errors) { admin.class.describe_errors }
 
-    it { expect(errors.size).to eq(162) }
+    it { expect(errors.size).to eq(168) }
     it { expect(errors[-184]).to eq(code: -184, description: 'Local: Queue full', name: '_QUEUE_FULL') }
     it { expect(errors[21]).to eq(code: 21, description: 'Broker: Invalid required acks value', name: 'INVALID_REQUIRED_ACKS') }
   end

data/spec/rdkafka/consumer_spec.rb

@@ -258,6 +258,95 @@ describe Rdkafka::Consumer do
     end
   end
 
+  describe "#seek_by" do
+    let(:topic) { "consume_test_topic" }
+    let(:partition) { 0 }
+    let(:offset) { 0 }
+
+    it "should raise an error when seeking fails" do
+      expect(Rdkafka::Bindings).to receive(:rd_kafka_seek).and_return(20)
+      expect {
+        consumer.seek_by(topic, partition, offset)
+      }.to raise_error Rdkafka::RdkafkaError
+    end
+
+    context "subscription" do
+      let(:timeout) { 1000 }
+
+      before do
+        consumer.subscribe(topic)
+
+        # 1. partitions are assigned
+        wait_for_assignment(consumer)
+        expect(consumer.assignment).not_to be_empty
+
+        # 2. eat unrelated messages
+        while(consumer.poll(timeout)) do; end
+      end
+      after { consumer.unsubscribe }
+
+      def send_one_message(val)
+        producer.produce(
+          topic:     topic,
+          payload:   "payload #{val}",
+          key:       "key 1",
+          partition: 0
+        ).wait
+      end
+
+      it "works when a partition is paused" do
+        # 3. get reference message
+        send_one_message(:a)
+        message1 = consumer.poll(timeout)
+        expect(message1&.payload).to eq "payload a"
+
+        # 4. pause the subscription
+        tpl = Rdkafka::Consumer::TopicPartitionList.new
+        tpl.add_topic(topic, 1)
+        consumer.pause(tpl)
+
+        # 5. seek by the previous message fields
+        consumer.seek_by(message1.topic, message1.partition, message1.offset)
+
+        # 6. resume the subscription
+        tpl = Rdkafka::Consumer::TopicPartitionList.new
+        tpl.add_topic(topic, 1)
+        consumer.resume(tpl)
+
+        # 7. ensure same message is read again
+        message2 = consumer.poll(timeout)
+
+        # This is needed because `enable.auto.offset.store` is true but when running in CI that
+        # is overloaded, offset store lags
+        sleep(2)
+
+        consumer.commit
+        expect(message1.offset).to eq message2.offset
+        expect(message1.payload).to eq message2.payload
+      end
+
+      it "allows skipping messages" do
+        # 3. send messages
+        send_one_message(:a)
+        send_one_message(:b)
+        send_one_message(:c)
+
+        # 4. get reference message
+        message = consumer.poll(timeout)
+        expect(message&.payload).to eq "payload a"
+
+        # 5. seek over one message
+        consumer.seek_by(message.topic, message.partition, message.offset + 2)
+
+        # 6. ensure that only one message is available
+        records = consumer.poll(timeout)
+        expect(records&.payload).to eq "payload c"
+        records = consumer.poll(timeout)
+        expect(records).to be_nil
+      end
+    end
+  end
+
   describe "#assign and #assignment" do
     it "should return an empty assignment if nothing is assigned" do
       expect(consumer.assignment).to be_empty

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: rdkafka
 version: !ruby/object:Gem::Version
-  version: 0.16.0
+  version: 0.17.0
 platform: ruby
 authors:
 - Thijs Cadier
@@ -36,7 +36,7 @@ cert_chain:
   AnG1dJU+yL2BK7vaVytLTstJME5mepSZ46qqIJXMuWob/YPDmVaBF39TDSG9e34s
   msG3BiCqgOgHAnL23+CN3Rt8MsuRfEtoTKpJVcCfoEoNHOkc
   -----END CERTIFICATE-----
-date: 2024-06-13 00:00:00.000000000 Z
+date: 2024-08-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
@@ -186,6 +186,7 @@ files:
 - README.md
 - Rakefile
 - certs/cert_chain.pem
+- dist/librdkafka_2.4.0.tar.gz
 - docker-compose.yml
 - ext/README.md
 - ext/Rakefile