Add compiler flags (#850)

Fixes #435

.github/actions/build_clio/action.yml

@@ -0,0 +1,36 @@
+name: Build clio
+description: Build clio in build directory
+inputs:
+  conan_profile:
+    description: Conan profile name
+    required: true
+    default: default
+  conan_cache_hit:
+    description: Whether conan cache has been downloaded
+    required: true
+runs:
+  using: composite
+  steps:
+    - name: Get number of threads on mac
+      id: mac_threads
+      if: ${{ runner.os == 'macOS' }}
+      shell: bash
+      run: echo "num=$(($(sysctl -n hw.logicalcpu) - 2))" >> $GITHUB_OUTPUT
+
+    - name: Get number of threads on Linux
+      id: linux_threads
+      if: ${{ runner.os == 'Linux' }}
+      shell: bash
+      run: echo "num=$(($(nproc) - 2))" >> $GITHUB_OUTPUT
+
+    - name: Build Clio
+      shell: bash
+      env:
+        BUILD_OPTION: "${{ inputs.conan_cache_hit == 'true' && 'missing' || '' }}"
+      run: |
+        mkdir -p build
+        cd build
+        threads_num=${{ steps.mac_threads.outputs.num || steps.linux_threads.outputs.num }}
+        conan install .. -of . -b $BUILD_OPTION -s build_type=Release -o clio:tests=True --profile ${{ inputs.conan_profile }}
+        cmake -DCMAKE_TOOLCHAIN_FILE:FILEPATH=build/generators/conan_toolchain.cmake -DCMAKE_BUILD_TYPE=Release .. -G Ninja
+        cmake --build . --parallel $threads_num

.github/actions/clang_format/action.yml

@@ -1,3 +1,5 @@
+name: Check format
+description: Check format using clang-format-11
 runs:
   using: composite
   steps:

.github/actions/git_common_ancestor/action.yml

@@ -0,0 +1,14 @@
+name: Git common ancestor
+description: Find the closest common commit
+outputs:
+  commit:
+    description: Hash of commit
+    value: ${{ steps.find_common_ancestor.outputs.commit }}
+runs:
+  using: composite
+  steps:
+    - name: Find common git ancestor
+      id: find_common_ancestor
+      shell: bash
+      run: |
+        echo "commit=$(git merge-base --fork-point origin/develop)" >> $GITHUB_OUTPUT

.github/actions/restore_cache/action.yml

@@ -0,0 +1,50 @@
+name: Restore cache
+description: Find and restores conan and ccache cache
+inputs:
+  conan_dir:
+    description: Path to .conan directory
+    required: true
+  ccache_dir:
+    description: Path to .ccache directory
+    required: true
+outputs:
+  conan_hash:
+    description: Hash to use as a part of conan cache key
+    value: ${{ steps.conan_hash.outputs.hash }}
+  conan_cache_hit:
+    description: True if conan cache has been downloaded
+    value: ${{ steps.conan_cache.outputs.cache-hit }}
+  ccache_cache_hit:
+    description: True if ccache cache has been downloaded
+    value: ${{ steps.ccache_cache.outputs.cache-hit }}
+runs:
+  using: composite
+  steps:
+    - name: Find common commit
+      id: git_common_ancestor
+      uses: ./.github/actions/git_common_ancestor
+
+    - name: Calculate conan hash
+      id: conan_hash
+      shell: bash
+      run: |
+        conan info . -j info.json
+        packages_info=$(cat info.json | jq '.[] | "\(.display_name): \(.id)"' | grep -v 'clio')
+        echo "$packages_info"
+        hash=$(echo "$packages_info" | shasum -a 256 | cut -d ' ' -f 1)
+        rm info.json
+        echo "hash=$hash" >> $GITHUB_OUTPUT
+
+    - name: Restore conan cache
+      uses: actions/cache/restore@v3
+      id: conan_cache
+      with:
+        path: ${{ inputs.conan_dir }}/data
+        key: clio-conan_data-${{ runner.os }}-develop-${{ steps.conan_hash.outputs.hash }}
+
+    - name: Restore ccache cache
+      uses: actions/cache/restore@v3
+      id: ccache_cache
+      with:
+        path: ${{ inputs.ccache_dir }}
+        key: clio-ccache-${{ runner.os }}-develop-${{ steps.git_common_ancestor.outputs.commit }}

.github/actions/save_cache/action.yml

@@ -0,0 +1,46 @@
+name: Save cache
+description: Save conan and ccache cache for develop branch
+inputs:
+  conan_dir:
+    description: Path to .conan directory
+    required: true
+  conan_hash:
+    description: Hash to use as a part of conan cache key
+    required: true
+  conan_cache_hit:
+    description: Whether conan cache has been downloaded
+    required: true
+  ccache_dir:
+    description: Path to .ccache directory
+    required: true
+  ccache_cache_hit:
+    description: Whether conan cache has been downloaded
+    required: true
+runs:
+  using: composite
+  steps:
+    - name: Find common commit
+      id: git_common_ancestor
+      uses: ./.github/actions/git_common_ancestor
+
+    - name: Cleanup conan directory from extra data
+      if: ${{ inputs.conan_cache_hit != 'true' }}
+      shell: bash
+      run: |
+        conan remove "*" -s -b -f
+
+    - name: Save conan cache
+      if: ${{ inputs.conan_cache_hit != 'true' }}
+      uses: actions/cache/save@v3
+      with:
+        path: ${{ inputs.conan_dir }}/data
+        key: clio-conan_data-${{ runner.os }}-develop-${{ inputs.conan_hash }}
+
+    - name: Save ccache cache
+      if: ${{ inputs.ccache_cache_hit != 'true' }}
+      uses: actions/cache/save@v3
+      with:
+        path: ${{ inputs.ccache_dir }}
+        key: clio-ccache-${{ runner.os }}-develop-${{ steps.git_common_ancestor.outputs.commit }}
+
+

.github/actions/setup_conan/action.yml

@@ -0,0 +1,55 @@
+name: Setup conan
+description: Setup conan profile and artifactory
+outputs:
+  conan_profile:
+    description: Created conan profile name
+    value: ${{ steps.conan_export_output.outputs.conan_profile }}
+runs:
+  using: composite
+  steps:
+    - name: On mac
+      if: ${{ runner.os == 'macOS' }}
+      shell: bash
+      env:
+        CONAN_PROFILE: clio_clang_14
+      id: conan_setup_mac
+      run: |
+        echo "Creating $CONAN_PROFILE conan profile";
+        clang_path="$(brew --prefix llvm@14)/bin/clang"
+        clang_cxx_path="$(brew --prefix llvm@14)/bin/clang++"
+        conan profile new $CONAN_PROFILE --detect --force
+        conan profile update settings.compiler=clang $CONAN_PROFILE
+        conan profile update settings.compiler.version=14 $CONAN_PROFILE
+        conan profile update settings.compiler.cppstd=20 $CONAN_PROFILE
+        conan profile update "conf.tools.build:compiler_executables={\"c\": \"$clang_path\", \"cpp\": \"$clang_cxx_path\"}" $CONAN_PROFILE
+        conan profile update env.CC="$clang_path" $CONAN_PROFILE
+        conan profile update env.CXX="$clang_cxx_path" $CONAN_PROFILE
+        echo "created_conan_profile=$CONAN_PROFILE" >> $GITHUB_OUTPUT
+
+    - name: On linux
+      if: ${{ runner.os == 'Linux' }}
+      shell: bash
+      id: conan_setup_linux
+      run: |
+        conan profile new default --detect
+        conan profile update settings.compiler.cppstd=20 default
+        conan profile update settings.compiler.libcxx=libstdc++11 default
+        echo "created_conan_profile=default" >> $GITHUB_OUTPUT
+
+    - name: Export output variable
+      shell: bash
+      id: conan_export_output
+      run: |
+        echo "conan_profile=${{ steps.conan_setup_mac.outputs.created_conan_profile || steps.conan_setup_linux.outputs.created_conan_profile }}" >> $GITHUB_OUTPUT
+
+    - name: Add conan-non-prod artifactory
+      shell: bash
+      run: |
+        if [[ -z $(conan remote list | grep conan-non-prod) ]]; then
+            echo "Adding conan-non-prod"
+            conan remote add --insert 0 conan-non-prod http://18.143.149.228:8081/artifactory/api/conan/conan-non-prod
+        else
+            echo "Conan-non-prod is available"
+        fi
+
+

.github/workflows/build.yml

@@ -1,9 +1,9 @@
 name: Build Clio
 on:
   push:
-    branches: [master, release/*, develop, develop-next]
+    branches: [master, release/*, develop]
   pull_request:
-    branches: [master, release/*, develop, develop-next]
+    branches: [master, release/*, develop]
   workflow_dispatch:
 
 jobs:
@@ -13,200 +13,142 @@ jobs:
     steps:
       - uses: actions/checkout@v3
       - name: Run clang-format
-        uses: ./.github/actions/lint
+        uses: ./.github/actions/clang_format
 
-  build_clio:
-    name: Build Clio
-    runs-on: [self-hosted, heavy]
+  build_mac:
+    name: Build macOS
     needs: lint
-    strategy:
-      fail-fast: false
-      matrix:
-        type:
-          - suffix: deb
-            image: rippleci/clio-dpkg-builder:2022-09-17
-            script: dpkg
-          - suffix: rpm
-            image: rippleci/clio-rpm-builder:2022-09-17
-            script: rpm
-
-    container:
-      image: ${{ matrix.type.image }}
+    runs-on: [self-hosted, macOS]
+    env:
+      CCACHE_DIR: ${{ github.workspace }}/.ccache
+      CONAN_USER_HOME: ${{ github.workspace }}
     steps:
       - uses: actions/checkout@v3
         with:
-          path: clio
-          fetch-depth: 0
+            fetch-depth: 0
 
-      - name: Clone Clio packaging repo
+      - name: Install packages
+        run: |
+          brew install llvm@14 pkg-config ninja bison cmake ccache jq
+
+      - name: Setup conan
+        uses: ./.github/actions/setup_conan
+        id: conan
+
+      - name: Restore cache
+        uses: ./.github/actions/restore_cache
+        id: restore_cache
+        with:
+          conan_dir: ${{ env.CONAN_USER_HOME }}/.conan
+          ccache_dir: ${{ env.CCACHE_DIR }}
+
+      - name: Build Clio
+        uses: ./.github/actions/build_clio
+        with:
+          conan_profile: ${{ steps.conan.outputs.conan_profile }}
+          conan_cache_hit: ${{ steps.restore_cache.outputs.conan_cache_hit }}
+
+      - name: Strip tests
+        run: strip build/clio_tests
+
+      - name: Upload clio_tests
+        uses: actions/upload-artifact@v3
+        with:
+          name: clio_tests_mac
+          path: build/clio_tests
+
+      - name: Save cache
+        uses: ./.github/actions/save_cache
+        with:
+          conan_dir: ${{ env.CONAN_USER_HOME }}/.conan
+          conan_hash: ${{ steps.restore_cache.outputs.conan_hash }}
+          conan_cache_hit: ${{ steps.restore_cache.outputs.conan_cache_hit }}
+          ccache_dir: ${{ env.CCACHE_DIR }}
+          ccache_cache_hit: ${{ steps.restore_cache.outputs.ccache_cache_hit }}
+
+  build_linux:
+    name: Build linux
+    needs: lint
+    runs-on: [self-hosted, Linux]
+    container:
+      image: conanio/gcc11:1.61.0
+      options: --user root
+    env:
+      CCACHE_DIR: /root/.ccache
+      CONAN_USER_HOME: /root/
+    steps:
+      - name: Get Clio
         uses: actions/checkout@v3
         with:
-          path: clio-packages
-          repository: XRPLF/clio-packages
-          ref: main
+          fetch-depth: 0
 
-      - name: Build
-        shell: bash
+      - name: Install packages
         run: |
-          export CLIO_ROOT=$(realpath clio)
-          if [ ${{ matrix.type.suffix }} == "rpm" ]; then
-            source /opt/rh/devtoolset-11/enable
-          fi
-          cmake -S clio-packages -B clio-packages/build -DCLIO_ROOT=$CLIO_ROOT
-          cmake --build clio-packages/build --parallel $(nproc)
-          cp ./clio-packages/build/clio-prefix/src/clio-build/clio_tests .
-          mv ./clio-packages/build/*.${{ matrix.type.suffix }} .
-      - name: Artifact packages
+          apt update -qq
+          apt install -y jq
+
+      - name: Install ccache
+        run: |
+          wget https://github.com/ccache/ccache/releases/download/v4.8.3/ccache-4.8.3-linux-x86_64.tar.xz
+          tar xf ./ccache-4.8.3-linux-x86_64.tar.xz
+          mv ./ccache-4.8.3-linux-x86_64/ccache /usr/bin/ccache
+
+      - name: Fix git permissions
+        run: git config --global --add safe.directory $PWD
+
+      - name: Setup conan
+        uses: ./.github/actions/setup_conan
+
+      - name: Restore cache
+        uses: ./.github/actions/restore_cache
+        id: restore_cache
+        with:
+          conan_dir: ${{ env.CONAN_USER_HOME }}/.conan
+          ccache_dir: ${{ env.CCACHE_DIR }}
+
+      - name: Build Clio
+        uses: ./.github/actions/build_clio
+        with:
+          conan_cache_hit: ${{ steps.restore_cache.outputs.conan_cache_hit }}
+
+      - name: Strip tests
+        run: strip build/clio_tests
+
+      - name: Upload clio_tests
         uses: actions/upload-artifact@v3
         with:
-          name: clio_${{ matrix.type.suffix }}_packages
-          path: ${{ github.workspace }}/*.${{ matrix.type.suffix }}
+          name: clio_tests_linux
+          path: build/clio_tests
 
-      - name: Artifact clio_tests
-        uses: actions/upload-artifact@v3
+      - name: Save cache
+        uses: ./.github/actions/save_cache
         with:
-          name: clio_tests-${{ matrix.type.suffix }}
-          path: ${{ github.workspace }}/clio_tests
+          conan_dir: ${{ env.CONAN_USER_HOME }}/.conan
+          conan_hash: ${{ steps.restore_cache.outputs.conan_hash }}
+          conan_cache_hit: ${{ steps.restore_cache.outputs.conan_cache_hit }}
+          ccache_dir: ${{ env.CCACHE_DIR }}
+          ccache_cache_hit: ${{ steps.restore_cache.outputs.ccache_cache_hit }}
 
-  build_dev:
-    name: Build on Mac/Clang14 and run tests
-    needs: lint
-    continue-on-error: false
+  test_mac:
+    needs: build_mac
     runs-on: [self-hosted, macOS]
-
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/download-artifact@v3
         with:
-          path: clio
-
-      - name: Check Boost cache
-        id: boost
-        uses: actions/cache@v3
-        with:
-          path: boost_1_77_0
-          key: ${{ runner.os }}-boost
-
-      - name: Build Boost
-        if: ${{ steps.boost.outputs.cache-hit != 'true' }}
+          name: clio_tests_mac
+      - name: Run clio_tests
         run: |
-          rm -rf boost_1_77_0.tar.gz boost_1_77_0 # cleanup if needed first
-          curl -s -fOJL "https://boostorg.jfrog.io/artifactory/main/release/1.77.0/source/boost_1_77_0.tar.gz"
-          tar zxf boost_1_77_0.tar.gz
-          cd boost_1_77_0
-          ./bootstrap.sh
-          ./b2 define=BOOST_ASIO_HAS_STD_INVOKE_RESULT cxxflags="-std=c++20"
-
-      - name: Install dependencies
-        run: |
-          brew install llvm@14 pkg-config protobuf openssl ninja cassandra-cpp-driver bison cmake
-
-      - name: Setup environment for llvm-14
-        run: |
-          export PATH="/usr/local/opt/llvm@14/bin:$PATH"
-          export LDFLAGS="-L/usr/local/opt/llvm@14/lib -L/usr/local/opt/llvm@14/lib/c++ -Wl,-rpath,/usr/local/opt/llvm@14/lib/c++"
-          export CPPFLAGS="-I/usr/local/opt/llvm@14/include"
-
-      - name: Build clio
-        run: |
-          export BOOST_ROOT=$(pwd)/boost_1_77_0
-          cd clio
-          cmake -B build -DCMAKE_C_COMPILER='/usr/local/opt/llvm@14/bin/clang' -DCMAKE_CXX_COMPILER='/usr/local/opt/llvm@14/bin/clang++'
-          if ! cmake --build build -j; then
-            echo '# 🔥🔥 MacOS AppleClang build failed!💥' >> $GITHUB_STEP_SUMMARY
-            exit 1
-          fi
-      - name: Run Test
-        run: |
-          cd clio/build
+          chmod +x ./clio_tests
           ./clio_tests --gtest_filter="-BackendCassandraBaseTest*:BackendCassandraTest*:BackendCassandraFactoryTestWithDB*"
 
-  test_clio:
-    name: Test Clio
-    runs-on: [self-hosted, Linux]
-    needs: build_clio
-    strategy:
-      fail-fast: false
-      matrix:
-        suffix: [rpm, deb]
+  test_linux:
+    needs: build_linux
+    runs-on: [self-hosted, x-heavy]
     steps:
-      - uses: actions/checkout@v3
-
-      - name: Get clio_tests artifact
-        uses: actions/download-artifact@v3
+      - uses: actions/download-artifact@v3
         with:
-          name: clio_tests-${{ matrix.suffix }}
-
-      - name: Run tests
-        timeout-minutes: 10
-        uses: ./.github/actions/test
-
-  code_coverage:
-    name: Build on Linux and code coverage
-    needs: lint
-    continue-on-error: false
-    runs-on: ubuntu-22.04
-
-    steps:
-      - uses: actions/checkout@v3
-        with:
-          path: clio
-
-      - name: Check Boost cache
-        id: boost
-        uses: actions/cache@v3
-        with:
-          path: boost
-          key: ${{ runner.os }}-boost
-
-      - name: Build boost
-        if: steps.boost.outputs.cache-hit != 'true'
+          name: clio_tests_linux
+      - name: Run clio_tests
         run: |
-          curl -s -OJL "https://boostorg.jfrog.io/artifactory/main/release/1.77.0/source/boost_1_77_0.tar.gz"
-          tar zxf boost_1_77_0.tar.gz
-          mv boost_1_77_0 boost
-          cd boost
-          ./bootstrap.sh
-          ./b2
-      - name: install deps
-        run: |
-          sudo apt-get -y install git pkg-config protobuf-compiler libprotobuf-dev libssl-dev wget build-essential doxygen bison flex autoconf clang-format gcovr
-      - name: Build clio
-        run: |
-          export BOOST_ROOT=$(pwd)/boost
-          cd clio
-          cmake -B build -DCODE_COVERAGE=on -DTEST_PARAMETER='--gtest_filter="-BackendCassandraBaseTest*:BackendCassandraTest*:BackendCassandraFactoryTestWithDB*"'
-          if ! cmake --build build -j$(nproc); then
-            echo '# 🔥Ubuntu build🔥 failed!💥' >> $GITHUB_STEP_SUMMARY
-            exit 1
-          fi
-          cd build
-          make clio_tests-ccov
-      - name: Code Coverage Summary Report
-        uses: irongut/CodeCoverageSummary@v1.2.0
-        with:
-          filename: clio/build/clio_tests-gcc-cov/out.xml
-          badge: true
-          output: both
-          format: markdown
-
-      - name: Save PR number and ccov report
-        run: |
-          mkdir -p ./UnitTestCoverage
-          echo ${{ github.event.number }} > ./UnitTestCoverage/NR
-          cp clio/build/clio_tests-gcc-cov/report.html ./UnitTestCoverage/report.html
-          cp code-coverage-results.md ./UnitTestCoverage/out.md
-          cat code-coverage-results.md > $GITHUB_STEP_SUMMARY
-      - name: Upload coverage reports to Codecov
-        uses: codecov/codecov-action@v3
-        with:
-          files: clio/build/clio_tests-gcc-cov/out.xml
-
-      - uses: actions/upload-artifact@v3
-        with:
-          name: UnitTestCoverage
-          path: UnitTestCoverage/
-
-      - uses: actions/upload-artifact@v3
-        with:
-          name: code_coverage_report
-          path: clio/build/clio_tests-gcc-cov/out.xml
+          chmod +x ./clio_tests
+          ./clio_tests --gtest_filter="-BackendCassandraBaseTest*:BackendCassandraTest*:BackendCassandraFactoryTestWithDB*"

.gitignore

@@ -1,6 +1,9 @@
 *clio*.log
-build*/
+/build*/
+.build
+.cache
 .vscode
 .python-version
+CMakeUserPresets.json
 config.json
 src/main/impl/Build.cpp

CMake/Ccache.cmake

@@ -0,0 +1,5 @@
+find_program (CCACHE_PATH "ccache")
+if (CCACHE_PATH)
+    set (CMAKE_CXX_COMPILER_LAUNCHER "${CCACHE_PATH}")
+    message (STATUS "Using ccache: ${CCACHE_PATH}")
+endif ()

CMake/CheckCompiler.cmake

@@ -0,0 +1,42 @@
+if (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
+  if (CMAKE_CXX_COMPILER_VERSION VERSION_LESS 14)
+    message (FATAL_ERROR "Clang 14+ required for building clio")
+  endif ()
+  set (is_clang TRUE)
+elseif (CMAKE_CXX_COMPILER_ID STREQUAL "AppleClang")
+  if (CMAKE_CXX_COMPILER_VERSION VERSION_LESS 14)
+    message (FATAL_ERROR "AppleClang 14+ required for building clio")
+  endif ()
+  set (is_appleclang TRUE)
+elseif (CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+  if (CMAKE_CXX_COMPILER_VERSION VERSION_LESS 11)
+    message (FATAL_ERROR "GCC 11+ required for building clio")
+  endif ()
+  set (is_gcc TRUE)
+else ()
+  message (FATAL_ERROR "Supported compilers: AppleClang 14+, Clang 14+, GCC 11+")
+endif ()
+
+if (san)
+  string (TOLOWER ${san} san)
+  set (SAN_FLAG "-fsanitize=${san}")
+  set (SAN_LIB "")
+  if (is_gcc)
+    if (san STREQUAL "address")
+      set (SAN_LIB "asan")
+    elseif (san STREQUAL "thread")
+      set (SAN_LIB "tsan")
+    elseif (san STREQUAL "memory")
+      set (SAN_LIB "msan")
+    elseif (san STREQUAL "undefined")
+      set (SAN_LIB "ubsan")
+    endif ()
+  endif ()
+  set (_saved_CRL ${CMAKE_REQUIRED_LIBRARIES})
+  set (CMAKE_REQUIRED_LIBRARIES "${SAN_FLAG};${SAN_LIB}")
+  CHECK_CXX_COMPILER_FLAG (${SAN_FLAG} COMPILER_SUPPORTS_SAN)
+  set (CMAKE_REQUIRED_LIBRARIES ${_saved_CRL})
+  if (NOT COMPILER_SUPPORTS_SAN)
+    message (FATAL_ERROR "${san} sanitizer does not seem to be supported by your compiler")
+  endif ()
+endif ()

CMake/ClioVersion.cmake

@@ -2,32 +2,38 @@
    write version to source
 #]===================================================================]
 
-find_package(Git REQUIRED)
+find_package (Git REQUIRED)
 
-set(GIT_COMMAND rev-parse --short HEAD)
-execute_process(COMMAND ${GIT_EXECUTABLE} ${GIT_COMMAND} OUTPUT_VARIABLE REV OUTPUT_STRIP_TRAILING_WHITESPACE)
+set (GIT_COMMAND rev-parse --short HEAD)
+execute_process (COMMAND ${GIT_EXECUTABLE} ${GIT_COMMAND} 
+  WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}
+  OUTPUT_VARIABLE REV OUTPUT_STRIP_TRAILING_WHITESPACE)
 
-set(GIT_COMMAND branch --show-current)
-execute_process(COMMAND ${GIT_EXECUTABLE} ${GIT_COMMAND} OUTPUT_VARIABLE BRANCH OUTPUT_STRIP_TRAILING_WHITESPACE)
+set (GIT_COMMAND branch --show-current)
+execute_process (COMMAND ${GIT_EXECUTABLE} ${GIT_COMMAND} 
+  WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}
+  OUTPUT_VARIABLE BRANCH OUTPUT_STRIP_TRAILING_WHITESPACE)
 
-if(BRANCH STREQUAL "")
-  set(BRANCH "dev")
-endif()
+if (BRANCH STREQUAL "")
+  set (BRANCH "dev")
+endif ()
 
-if(NOT (BRANCH MATCHES master OR BRANCH MATCHES release/*)) # for develop and any other branch name YYYYMMDDHMS-<branch>-<git-ref>
-  execute_process(COMMAND date +%Y%m%d%H%M%S OUTPUT_VARIABLE DATE OUTPUT_STRIP_TRAILING_WHITESPACE)
-  set(VERSION "${DATE}-${BRANCH}-${REV}")
-else()
-  set(GIT_COMMAND describe --tags)
-  execute_process(COMMAND ${GIT_EXECUTABLE} ${GIT_COMMAND} OUTPUT_VARIABLE TAG_VERSION OUTPUT_STRIP_TRAILING_WHITESPACE)
-  set(VERSION "${TAG_VERSION}-${REV}")
-endif()
+if (NOT (BRANCH MATCHES master OR BRANCH MATCHES release/*)) # for develop and any other branch name YYYYMMDDHMS-<branch>-<git-rev>
+  execute_process (COMMAND date +%Y%m%d%H%M%S OUTPUT_VARIABLE DATE OUTPUT_STRIP_TRAILING_WHITESPACE)
+  set (VERSION "${DATE}-${BRANCH}-${REV}")
+else ()
+  set (GIT_COMMAND describe --tags)
+  execute_process (COMMAND ${GIT_EXECUTABLE} ${GIT_COMMAND} 
+    WORKING_DIRECTORY ${CMAKE_SOURCE_DIR}
+    OUTPUT_VARIABLE TAG_VERSION OUTPUT_STRIP_TRAILING_WHITESPACE)
+  set (VERSION "${TAG_VERSION}-${REV}")
+endif ()
 
-if(CMAKE_BUILD_TYPE MATCHES Debug)
-  set(VERSION "${VERSION}+DEBUG")
-endif()
+if (CMAKE_BUILD_TYPE MATCHES Debug)
+  set (VERSION "${VERSION}+DEBUG")
+endif ()
 
-message(STATUS "Build version: ${VERSION}")
-set(clio_version "${VERSION}")
+message (STATUS "Build version: ${VERSION}")
+set (clio_version "${VERSION}")
 
-configure_file(CMake/Build.cpp.in ${CMAKE_SOURCE_DIR}/src/main/impl/Build.cpp)
+configure_file (CMake/Build.cpp.in ${CMAKE_SOURCE_DIR}/src/main/impl/Build.cpp)

CMake/Coverage.cmake

@@ -1,42 +1,45 @@
-# call add_converage(module_name) to add coverage targets for the given module
-function(add_converage module)
-  if("${CMAKE_C_COMPILER_ID}" MATCHES "(Apple)?[Cc]lang"
+# call add_coverage(module_name) to add coverage targets for the given module
+function (add_coverage module)
+  if ("${CMAKE_C_COMPILER_ID}" MATCHES "(Apple)?[Cc]lang"
      OR "${CMAKE_CXX_COMPILER_ID}" MATCHES "(Apple)?[Cc]lang")
-    message("[Coverage] Building with llvm Code Coverage Tools")
+    message ("[Coverage] Building with llvm Code Coverage Tools")
     # Using llvm gcov ; llvm install by xcode
-    set(LLVM_COV_PATH /Library/Developer/CommandLineTools/usr/bin)
-    if(NOT EXISTS ${LLVM_COV_PATH}/llvm-cov)
-      message(FATAL_ERROR "llvm-cov not found! Aborting.")
-    endif()
+    set (LLVM_COV_PATH /Library/Developer/CommandLineTools/usr/bin)
+    if (NOT EXISTS ${LLVM_COV_PATH}/llvm-cov)
+      message (FATAL_ERROR "llvm-cov not found! Aborting.")
+    endif ()
 
     # set Flags
-    target_compile_options(${module} PRIVATE -fprofile-instr-generate
-                                             -fcoverage-mapping)
-    target_link_options(${module} PUBLIC -fprofile-instr-generate
-                        -fcoverage-mapping)
+    target_compile_options (${module} PRIVATE 
+      -fprofile-instr-generate
+      -fcoverage-mapping)
+    
+    target_link_options (${module} PUBLIC 
+      -fprofile-instr-generate
+      -fcoverage-mapping)
 
-    target_compile_options(clio PRIVATE -fprofile-instr-generate
-                                        -fcoverage-mapping)
-    target_link_options(clio PUBLIC -fprofile-instr-generate
-                        -fcoverage-mapping)
+    target_compile_options (clio PRIVATE 
+      -fprofile-instr-generate
+      -fcoverage-mapping)
+    
+    target_link_options (clio PUBLIC 
+      -fprofile-instr-generate
+      -fcoverage-mapping)
 
     # llvm-cov
-    add_custom_target(
-      ${module}-ccov-preprocessing
+    add_custom_target (${module}-ccov-preprocessing
       COMMAND LLVM_PROFILE_FILE=${module}.profraw $<TARGET_FILE:${module}>
       COMMAND ${LLVM_COV_PATH}/llvm-profdata merge -sparse ${module}.profraw -o
               ${module}.profdata
       DEPENDS ${module})
 
-    add_custom_target(
-      ${module}-ccov-show
+    add_custom_target (${module}-ccov-show
       COMMAND ${LLVM_COV_PATH}/llvm-cov show $<TARGET_FILE:${module}>
               -instr-profile=${module}.profdata -show-line-counts-or-regions
       DEPENDS ${module}-ccov-preprocessing)
 
     # add summary for CI parse
-    add_custom_target(
-      ${module}-ccov-report
+    add_custom_target (${module}-ccov-report
       COMMAND
         ${LLVM_COV_PATH}/llvm-cov report $<TARGET_FILE:${module}>
         -instr-profile=${module}.profdata
@@ -45,8 +48,7 @@ function(add_converage module)
       DEPENDS ${module}-ccov-preprocessing)
 
     # exclude libs and unittests self
-    add_custom_target(
-      ${module}-ccov
+    add_custom_target (${module}-ccov
       COMMAND
         ${LLVM_COV_PATH}/llvm-cov show $<TARGET_FILE:${module}>
         -instr-profile=${module}.profdata -show-line-counts-or-regions
@@ -54,38 +56,36 @@ function(add_converage module)
         -ignore-filename-regex=".*_makefiles|.*unittests|.*_deps" > /dev/null 2>&1
       DEPENDS ${module}-ccov-preprocessing)
 
-    add_custom_command(
+    add_custom_command (
       TARGET ${module}-ccov
       POST_BUILD
       COMMENT
         "Open ${module}-llvm-cov/index.html in your browser to view the coverage report."
     )
-  elseif("${CMAKE_C_COMPILER_ID}" MATCHES "GNU" OR "${CMAKE_CXX_COMPILER_ID}"
-                                                   MATCHES "GNU")
-    message("[Coverage] Building with Gcc Code Coverage Tools")
+  elseif ("${CMAKE_C_COMPILER_ID}" MATCHES "GNU" OR "${CMAKE_CXX_COMPILER_ID}" MATCHES "GNU")
+    message ("[Coverage] Building with Gcc Code Coverage Tools")
 
-    find_program(GCOV_PATH gcov)
-    if(NOT GCOV_PATH)
-      message(FATAL_ERROR "gcov not found! Aborting...")
-    endif() # NOT GCOV_PATH
-    find_program(GCOVR_PATH gcovr)
-    if(NOT GCOVR_PATH)
-      message(FATAL_ERROR "gcovr not found! Aborting...")
-    endif() # NOT GCOVR_PATH
+    find_program (GCOV_PATH gcov)
+    if (NOT GCOV_PATH)
+      message (FATAL_ERROR "gcov not found! Aborting...")
+    endif () # NOT GCOV_PATH
+    find_program (GCOVR_PATH gcovr)
+    if (NOT GCOVR_PATH)
+      message (FATAL_ERROR "gcovr not found! Aborting...")
+    endif () # NOT GCOVR_PATH
 
-    set(COV_OUTPUT_PATH ${module}-gcc-cov)
-    target_compile_options(${module} PRIVATE -fprofile-arcs -ftest-coverage
+    set (COV_OUTPUT_PATH ${module}-gcc-cov)
+    target_compile_options (${module} PRIVATE -fprofile-arcs -ftest-coverage
                                              -fPIC)
-    target_link_libraries(${module} PRIVATE gcov)
+    target_link_libraries (${module} PRIVATE gcov)
 
-    target_compile_options(clio PRIVATE -fprofile-arcs -ftest-coverage
+    target_compile_options (clio PRIVATE -fprofile-arcs -ftest-coverage
                                              -fPIC)
-    target_link_libraries(clio PRIVATE gcov)
+    target_link_libraries (clio PRIVATE gcov)
     # this target is used for CI as well generate the summary out.xml will send
     # to github action to generate markdown, we can paste it to comments or
     # readme
-    add_custom_target(
-      ${module}-ccov
+    add_custom_target (${module}-ccov
       COMMAND ${module} ${TEST_PARAMETER}
       COMMAND rm -rf ${COV_OUTPUT_PATH}
       COMMAND mkdir ${COV_OUTPUT_PATH}
@@ -102,8 +102,7 @@ function(add_converage module)
       COMMENT "Running gcovr to produce Cobertura code coverage report.")
 
     # generate the detail report
-    add_custom_target(
-      ${module}-ccov-report
+    add_custom_target (${module}-ccov-report
       COMMAND ${module} ${TEST_PARAMETER}
       COMMAND rm -rf ${COV_OUTPUT_PATH}
       COMMAND mkdir ${COV_OUTPUT_PATH}
@@ -114,13 +113,13 @@ function(add_converage module)
         --exclude='${PROJECT_BINARY_DIR}/'
       WORKING_DIRECTORY ${PROJECT_BINARY_DIR}
       COMMENT "Running gcovr to produce Cobertura code coverage report.")
-    add_custom_command(
+    add_custom_command (
       TARGET ${module}-ccov-report
       POST_BUILD
       COMMENT
         "Open ${COV_OUTPUT_PATH}/index.html in your browser to view the coverage report."
     )
-  else()
-    message(FATAL_ERROR "Complier not support yet")
-  endif()
-endfunction()
+  else ()
+    message (FATAL_ERROR "Complier not support yet")
+  endif ()
+endfunction ()

CMake/Docs.cmake

@@ -0,0 +1,11 @@
+find_package (Doxygen REQUIRED)
+
+set (DOXYGEN_IN ${CMAKE_CURRENT_SOURCE_DIR}/Doxyfile)
+set (DOXYGEN_OUT ${CMAKE_CURRENT_BINARY_DIR}/Doxyfile)
+
+configure_file (${DOXYGEN_IN} ${DOXYGEN_OUT} @ONLY)
+add_custom_target (docs
+    COMMAND ${DOXYGEN_EXECUTABLE} ${DOXYGEN_OUT}
+    WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}
+    COMMENT "Generating API documentation with Doxygen"
+    VERBATIM)

CMake/Settings.cmake

@@ -0,0 +1,39 @@
+set(COMPILER_FLAGS
+  -Wall
+  -Wcast-align
+  -Wdouble-promotion
+  -Wextra
+  -Werror
+  -Wformat=2
+  -Wimplicit-fallthrough
+  -Wmisleading-indentation
+  -Wno-narrowing
+  -Wno-deprecated-declarations
+  -Wno-dangling-else
+  -Wno-unused-but-set-variable
+  -Wnon-virtual-dtor
+  -Wnull-dereference
+  -Wold-style-cast
+  -pedantic
+  -Wpedantic
+  -Wunused
+)
+
+if (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
+  list(APPEND COMPILER_FLAGS
+    -Wshadow # gcc is to aggressive with shadowing https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78147
+  )
+endif ()
+
+if (CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+  list(APPEND COMPILER_FLAGS
+    -Wduplicated-branches
+    -Wduplicated-cond
+    -Wlogical-op
+    -Wuseless-cast
+  )
+endif ()
+
+# See https://github.com/cpp-best-practices/cppbestpractices/blob/master/02-Use_the_Tools_Available.md#gcc--clang for the flags description
+
+target_compile_options (clio PUBLIC ${COMPILER_FLAGS})

CMake/SourceLocation.cmake

@@ -0,0 +1,11 @@
+include (CheckIncludeFileCXX)
+
+check_include_file_cxx ("source_location" SOURCE_LOCATION_AVAILABLE)
+if (SOURCE_LOCATION_AVAILABLE) 
+    target_compile_definitions (clio PUBLIC "HAS_SOURCE_LOCATION")
+endif ()
+
+check_include_file_cxx ("experimental/source_location" EXPERIMENTAL_SOURCE_LOCATION_AVAILABLE)
+if (EXPERIMENTAL_SOURCE_LOCATION_AVAILABLE)     
+    target_compile_definitions (clio PUBLIC "HAS_EXPERIMENTAL_SOURCE_LOCATION")
+endif ()

CMake/deps/Boost.cmake

@@ -1,6 +1,11 @@
-set(Boost_USE_STATIC_LIBS ON)
-set(Boost_USE_STATIC_RUNTIME ON)
+set (Boost_USE_STATIC_LIBS ON)
+set (Boost_USE_STATIC_RUNTIME ON)
 
-find_package(Boost 1.75 COMPONENTS filesystem log_setup log thread system REQUIRED)
-
-target_link_libraries(clio PUBLIC ${Boost_LIBRARIES})
+find_package (Boost 1.82 REQUIRED
+  COMPONENTS
+    program_options
+    coroutine
+    system
+    log
+    log_setup
+)

CMake/deps/OpenSSL.cmake

@@ -0,0 +1,5 @@
+find_package (OpenSSL 1.1.1 REQUIRED)
+
+set_target_properties (OpenSSL::SSL PROPERTIES
+  INTERFACE_COMPILE_DEFINITIONS OPENSSL_NO_SSL2
+)

CMake/deps/Remove-bitset-operator.patch

@@ -1,24 +0,0 @@
-From 5cd9d09d960fa489a0c4379880cd7615b1c16e55 Mon Sep 17 00:00:00 2001
-From: CJ Cobb <ccobb@ripple.com>
-Date: Wed, 10 Aug 2022 12:30:01 -0400
-Subject: [PATCH] Remove bitset operator !=
-
----
- src/ripple/protocol/Feature.h | 1 -
- 1 file changed, 1 deletion(-)
-
-diff --git a/src/ripple/protocol/Feature.h b/src/ripple/protocol/Feature.h
-index b3ecb099b..6424be411 100644
---- a/src/ripple/protocol/Feature.h
-+++ b/src/ripple/protocol/Feature.h
-@@ -126,7 +126,6 @@ class FeatureBitset : private std::bitset<detail::numFeatures>
- public:
-     using base::bitset;
-     using base::operator==;
--    using base::operator!=;
- 
-     using base::all;
-     using base::any;
--- 
-2.32.0
-

CMake/deps/SourceLocation.cmake

@@ -1,11 +0,0 @@
-include(CheckIncludeFileCXX)
-
-check_include_file_cxx("source_location" SOURCE_LOCATION_AVAILABLE)
-if(SOURCE_LOCATION_AVAILABLE) 
-    target_compile_definitions(clio PUBLIC "HAS_SOURCE_LOCATION")
-endif()
-
-check_include_file_cxx("experimental/source_location" EXPERIMENTAL_SOURCE_LOCATION_AVAILABLE)
-if(EXPERIMENTAL_SOURCE_LOCATION_AVAILABLE)     
-    target_compile_definitions(clio PUBLIC "HAS_EXPERIMENTAL_SOURCE_LOCATION")
-endif()

CMake/deps/Threads.cmake

@@ -0,0 +1,2 @@
+set (THREADS_PREFER_PTHREAD_FLAG ON)
+find_package (Threads)

CMake/deps/cassandra.cmake

@@ -1,153 +1 @@
-find_package(ZLIB REQUIRED)
-
-find_library(cassandra NAMES cassandra)
-if(NOT cassandra)
-    message("System installed Cassandra cpp driver not found. Will build")
-    find_library(zlib NAMES zlib1g-dev zlib-devel zlib z)
-    if(NOT zlib)
-        message("zlib not found. will build")
-        add_library(zlib STATIC IMPORTED GLOBAL)
-        ExternalProject_Add(zlib_src
-            PREFIX ${nih_cache_path}
-            GIT_REPOSITORY https://github.com/madler/zlib.git
-            GIT_TAG v1.2.12
-            INSTALL_COMMAND ""
-            BUILD_BYPRODUCTS <BINARY_DIR>/${CMAKE_STATIC_LIBRARY_PREFIX}z.a
-            )
-        ExternalProject_Get_Property (zlib_src SOURCE_DIR)
-        ExternalProject_Get_Property (zlib_src BINARY_DIR)
-        set (zlib_src_SOURCE_DIR "${SOURCE_DIR}")
-        file (MAKE_DIRECTORY ${zlib_src_SOURCE_DIR}/include)
-        set_target_properties (zlib PROPERTIES
-            IMPORTED_LOCATION
-            ${BINARY_DIR}/${CMAKE_STATIC_LIBRARY_PREFIX}z.a
-            INTERFACE_INCLUDE_DIRECTORIES
-            ${SOURCE_DIR}/include)
-        add_dependencies(zlib zlib_src)
-        file(TO_CMAKE_PATH "${zlib_src_SOURCE_DIR}" zlib_src_SOURCE_DIR)
-    endif()
-    find_library(krb5 NAMES krb5-dev libkrb5-dev)
-    if(NOT krb5)
-        message("krb5 not found. will build")
-        add_library(krb5 STATIC IMPORTED GLOBAL)
-        ExternalProject_Add(krb5_src
-            PREFIX ${nih_cache_path}
-            GIT_REPOSITORY https://github.com/krb5/krb5.git
-            GIT_TAG krb5-1.20
-            UPDATE_COMMAND ""
-            CONFIGURE_COMMAND autoreconf src && CFLAGS=-fcommon ./src/configure --enable-static --disable-shared
-            BUILD_IN_SOURCE 1
-            BUILD_COMMAND make
-            INSTALL_COMMAND ""
-            BUILD_BYPRODUCTS <SOURCE_DIR>/lib/${CMAKE_STATIC_LIBRARY_PREFIX}krb5.a
-            )
-        message(${ep_lib_prefix}/krb5.a)
-        message(${CMAKE_STATIC_LIBRARY_PREFIX}krb5.a)
-        ExternalProject_Get_Property (krb5_src SOURCE_DIR)
-        ExternalProject_Get_Property (krb5_src BINARY_DIR)
-        set (krb5_src_SOURCE_DIR "${SOURCE_DIR}")
-        file (MAKE_DIRECTORY ${krb5_src_SOURCE_DIR}/include)
-        set_target_properties (krb5 PROPERTIES
-            IMPORTED_LOCATION
-	    ${SOURCE_DIR}/lib/${CMAKE_STATIC_LIBRARY_PREFIX}krb5.a
-            INTERFACE_INCLUDE_DIRECTORIES
-            ${SOURCE_DIR}/include)
-        add_dependencies(krb5 krb5_src)
-        file(TO_CMAKE_PATH "${krb5_src_SOURCE_DIR}" krb5_src_SOURCE_DIR)
-    endif()
-
-
-    find_library(libuv1 NAMES uv1 libuv1 liubuv1-dev libuv1:amd64)
-
-
-    if(NOT libuv1)
-        message("libuv1 not found, will build")
-        add_library(libuv1 STATIC IMPORTED GLOBAL)
-        ExternalProject_Add(libuv_src
-            PREFIX ${nih_cache_path}
-            GIT_REPOSITORY https://github.com/libuv/libuv.git
-            GIT_TAG v1.44.1
-            INSTALL_COMMAND ""
-            BUILD_BYPRODUCTS <BINARY_DIR>/${CMAKE_STATIC_LIBRARY_PREFIX}uv_a.a
-            )
-
-        ExternalProject_Get_Property (libuv_src SOURCE_DIR)
-        ExternalProject_Get_Property (libuv_src BINARY_DIR)
-        set (libuv_src_SOURCE_DIR "${SOURCE_DIR}")
-        file (MAKE_DIRECTORY ${libuv_src_SOURCE_DIR}/include)
-
-        set_target_properties (libuv1 PROPERTIES
-            IMPORTED_LOCATION
-            ${BINARY_DIR}/${CMAKE_STATIC_LIBRARY_PREFIX}uv_a.a
-            INTERFACE_INCLUDE_DIRECTORIES
-            ${SOURCE_DIR}/include)
-        add_dependencies(libuv1 libuv_src)
-
-        file(TO_CMAKE_PATH "${libuv_src_SOURCE_DIR}" libuv_src_SOURCE_DIR)
-    endif()
-    add_library (cassandra STATIC IMPORTED GLOBAL)
-    ExternalProject_Add(cassandra_src
-        PREFIX ${nih_cache_path}
-        GIT_REPOSITORY https://github.com/datastax/cpp-driver.git
-        GIT_TAG 2.16.2
-        CMAKE_ARGS
-        -DLIBUV_ROOT_DIR=${BINARY_DIR}
-        -DLIBUV_INCLUDE_DIR=${SOURCE_DIR}/include
-        -DCASS_BUILD_STATIC=ON
-        -DCASS_BUILD_SHARED=OFF
-        INSTALL_COMMAND ""
-        BUILD_BYPRODUCTS <BINARY_DIR>/${CMAKE_STATIC_LIBRARY_PREFIX}cassandra_static.a
-        )
-
-    ExternalProject_Get_Property (cassandra_src SOURCE_DIR)
-    ExternalProject_Get_Property (cassandra_src BINARY_DIR)
-    set (cassandra_src_SOURCE_DIR "${SOURCE_DIR}")
-    file (MAKE_DIRECTORY ${cassandra_src_SOURCE_DIR}/include)
-
-    set_target_properties (cassandra PROPERTIES
-        IMPORTED_LOCATION
-        ${BINARY_DIR}/${CMAKE_STATIC_LIBRARY_PREFIX}cassandra_static.a
-        INTERFACE_INCLUDE_DIRECTORIES
-        ${SOURCE_DIR}/include)
-    message("cass dirs")
-    message(${BINARY_DIR})
-    message(${SOURCE_DIR})
-    message(${BINARY_DIR}/${CMAKE_STATIC_LIBRARY_PREFIX}cassandra_static.a)
-    add_dependencies(cassandra cassandra_src)
-
-    if(NOT libuv1)
-        ExternalProject_Add_StepDependencies(cassandra_src build libuv1)
-        target_link_libraries(cassandra INTERFACE libuv1)
-    else()
-        target_link_libraries(cassandra INTERFACE ${libuv1})
-    endif()
-    if(NOT krb5)
-
-        ExternalProject_Add_StepDependencies(cassandra_src build krb5)
-        target_link_libraries(cassandra INTERFACE krb5)
-    else()
-        target_link_libraries(cassandra INTERFACE ${krb5})
-    endif()
-
-    if(NOT zlib)
-        ExternalProject_Add_StepDependencies(cassandra_src build zlib)
-        target_link_libraries(cassandra INTERFACE zlib)
-    else()
-        target_link_libraries(cassandra INTERFACE ${zlib})
-    endif()
-	set(OPENSSL_USE_STATIC_LIBS TRUE)
-	find_package(OpenSSL REQUIRED)
-	target_link_libraries(cassandra INTERFACE OpenSSL::SSL)
-
-    file(TO_CMAKE_PATH "${cassandra_src_SOURCE_DIR}" cassandra_src_SOURCE_DIR)
-    target_link_libraries(clio PUBLIC cassandra)
-else()
-    message("Found system installed cassandra cpp driver")
-    message(${cassandra})
-    find_path(cassandra_includes NAMES cassandra.h REQUIRED)
-    message(${cassandra_includes})
-    get_filename_component(CASSANDRA_HEADER ${cassandra_includes}/cassandra.h REALPATH)
-    get_filename_component(CASSANDRA_HEADER_DIR ${CASSANDRA_HEADER} DIRECTORY)
-    target_link_libraries (clio PUBLIC ${cassandra})
-    target_include_directories(clio PUBLIC ${CASSANDRA_HEADER_DIR})
-endif()
+find_package (cassandra-cpp-driver REQUIRED)

CMake/deps/gtest.cmake

@@ -1,22 +1,4 @@
-FetchContent_Declare(
-  googletest
-  URL https://github.com/google/googletest/archive/609281088cfefc76f9d0ce82e1ff6c30cc3591e5.zip
-)
+find_package (GTest REQUIRED)
 
-FetchContent_GetProperties(googletest)
-
-if(NOT googletest_POPULATED)
-  FetchContent_Populate(googletest)
-  add_subdirectory(${googletest_SOURCE_DIR} ${googletest_BINARY_DIR} EXCLUDE_FROM_ALL)
-endif()
-
-target_link_libraries(clio_tests PUBLIC clio gmock_main)
-target_include_directories(clio_tests PRIVATE unittests)
-
-enable_testing()
-
-include(GoogleTest)
-
-#increase timeout for tests discovery to 10 seconds, by default it is 5s. As more unittests added, we start to hit this issue
-#https://github.com/google/googletest/issues/3475
-gtest_discover_tests(clio_tests DISCOVERY_TIMEOUT 10)
+enable_testing ()
+include (GoogleTest)

CMake/deps/libfmt.cmake

@@ -1,14 +1 @@
-FetchContent_Declare(
-  libfmt
-  URL https://github.com/fmtlib/fmt/releases/download/9.1.0/fmt-9.1.0.zip
-)
-
-FetchContent_GetProperties(libfmt)
-
-if(NOT libfmt_POPULATED)
-  FetchContent_Populate(libfmt)
-  add_subdirectory(${libfmt_SOURCE_DIR} ${libfmt_BINARY_DIR} EXCLUDE_FROM_ALL)
-endif()
-
-target_link_libraries(clio PUBLIC fmt)
-
+find_package (fmt REQUIRED)

CMake/deps/libxrpl.cmake

@@ -0,0 +1 @@
+find_package (xrpl REQUIRED)

CMake/deps/rippled.cmake

@@ -1,20 +0,0 @@
-set(RIPPLED_REPO "https://github.com/ripple/rippled.git")
-set(RIPPLED_BRANCH "1.9.2")
-set(NIH_CACHE_ROOT "${CMAKE_CURRENT_BINARY_DIR}" CACHE INTERNAL "")
-set(patch_command ! grep operator!= src/ripple/protocol/Feature.h || git apply < ${CMAKE_CURRENT_SOURCE_DIR}/CMake/deps/Remove-bitset-operator.patch)
-message(STATUS "Cloning ${RIPPLED_REPO} branch ${RIPPLED_BRANCH}")
-FetchContent_Declare(rippled
-  GIT_REPOSITORY "${RIPPLED_REPO}"
-  GIT_TAG "${RIPPLED_BRANCH}"
-  GIT_SHALLOW ON
-  PATCH_COMMAND "${patch_command}"
-)
-
-FetchContent_GetProperties(rippled)
-if(NOT rippled_POPULATED)
-  FetchContent_Populate(rippled)
-  add_subdirectory(${rippled_SOURCE_DIR} ${rippled_BINARY_DIR} EXCLUDE_FROM_ALL)
-endif()
-
-target_link_libraries(clio PUBLIC xrpl_core grpc_pbufs)
-target_include_directories(clio PUBLIC ${rippled_SOURCE_DIR}/src ) # TODO: Seems like this shouldn't be needed?

CMake/install/install.cmake

@@ -1,16 +1,14 @@
-set(CLIO_INSTALL_DIR "/opt/clio")
-set(CMAKE_INSTALL_PREFIX ${CLIO_INSTALL_DIR})
+set (CLIO_INSTALL_DIR "/opt/clio")
+set (CMAKE_INSTALL_PREFIX ${CLIO_INSTALL_DIR})
 
-install(TARGETS clio_server DESTINATION bin)
-# install(TARGETS clio_tests DESTINATION bin) # NOTE: Do we want to install the tests?
+install (TARGETS clio_server DESTINATION bin)
 
-#install(FILES example-config.json DESTINATION etc RENAME config.json)
-file(READ example-config.json config)
-string(REGEX REPLACE "./clio_log" "/var/log/clio/" config "${config}")
-file(WRITE ${CMAKE_BINARY_DIR}/install-config.json "${config}")
-install(FILES ${CMAKE_BINARY_DIR}/install-config.json DESTINATION etc RENAME config.json)
+file (READ example-config.json config)
+string (REGEX REPLACE "./clio_log" "/var/log/clio/" config "${config}")
+file (WRITE ${CMAKE_BINARY_DIR}/install-config.json "${config}")
+install (FILES ${CMAKE_BINARY_DIR}/install-config.json DESTINATION etc RENAME config.json)
 
-configure_file("${CMAKE_SOURCE_DIR}/CMake/install/clio.service.in" "${CMAKE_BINARY_DIR}/clio.service")
+configure_file ("${CMAKE_SOURCE_DIR}/CMake/install/clio.service.in" "${CMAKE_BINARY_DIR}/clio.service")
 
-install(FILES "${CMAKE_BINARY_DIR}/clio.service" DESTINATION /lib/systemd/system)
+install (FILES "${CMAKE_BINARY_DIR}/clio.service" DESTINATION /lib/systemd/system)
 

CMake/settings.cmake

@@ -1,6 +0,0 @@
-target_compile_options(clio  
-  PUBLIC -Wall
-         -Werror 
-         -Wno-narrowing
-         -Wno-deprecated-declarations 
-         -Wno-dangling-else)

CMakeLists.txt

@@ -1,56 +1,87 @@
 cmake_minimum_required(VERSION 3.16.3)
-
 project(clio)
 
-if(CMAKE_CXX_COMPILER_VERSION VERSION_LESS 11)
-  message(FATAL_ERROR "GCC 11+ required for building clio")
-endif()
+# ========================================================================== #
+#                                   Options                                  #
+# ========================================================================== #
+option (verbose   "Verbose build"                                     FALSE)
+option (tests     "Build tests"                                       FALSE)
+option (docs      "Generate doxygen docs"                             FALSE)
+option (coverage  "Build test coverage report"                        FALSE)
+option (packaging "Create distribution packages"                      FALSE)
+# ========================================================================== #
+set (san "" CACHE STRING "Add sanitizer instrumentation")
+set (CMAKE_EXPORT_COMPILE_COMMANDS TRUE)
+set_property (CACHE san PROPERTY STRINGS ";undefined;memory;address;thread")
+# ========================================================================== #
 
-option(BUILD_TESTS "Build tests" TRUE)
+# Include required modules
+include (CMake/Ccache.cmake)
+include (CheckCXXCompilerFlag)
 
-option(VERBOSE "Verbose build" TRUE)
-if(VERBOSE)
-  set(CMAKE_VERBOSE_MAKEFILE TRUE)
-  set(FETCHCONTENT_QUIET FALSE CACHE STRING "Verbose FetchContent()")
-endif()
+if (verbose)
+  set (CMAKE_VERBOSE_MAKEFILE TRUE)
+endif ()
 
+if (packaging)
+  add_definitions (-DPKG=1)
+endif ()
 
-if(PACKAGING)
-  add_definitions(-DPKG=1)
-endif()
+add_library (clio)
 
-#c++20 removed std::result_of but boost 1.75 is still using it.
-add_definitions(-DBOOST_ASIO_HAS_STD_INVOKE_RESULT=1)
+# Clio tweaks and checks
+include (CMake/CheckCompiler.cmake)
+include (CMake/Settings.cmake)
+include (CMake/ClioVersion.cmake)
+include (CMake/SourceLocation.cmake)
 
-add_library(clio)
-target_compile_features(clio PUBLIC cxx_std_20)
-target_include_directories(clio PUBLIC src)
+# Clio deps
+include (CMake/deps/libxrpl.cmake)
+include (CMake/deps/Boost.cmake)
+include (CMake/deps/OpenSSL.cmake)
+include (CMake/deps/Threads.cmake)
+include (CMake/deps/libfmt.cmake)
+include (CMake/deps/cassandra.cmake)
 
-include(FetchContent)
-include(ExternalProject)
-include(CMake/settings.cmake)
-include(CMake/ClioVersion.cmake)
-include(CMake/deps/rippled.cmake)
-include(CMake/deps/libfmt.cmake)
-include(CMake/deps/Boost.cmake)
-include(CMake/deps/cassandra.cmake)
-include(CMake/deps/SourceLocation.cmake)
+# TODO: Include directory will be wrong when installed.
+target_include_directories (clio PUBLIC src)
+target_compile_features (clio PUBLIC cxx_std_20)
 
-target_sources(clio PRIVATE
+target_link_libraries (clio
+  PUBLIC Boost::boost
+  PUBLIC Boost::coroutine
+  PUBLIC Boost::program_options
+  PUBLIC Boost::system
+  PUBLIC Boost::log
+  PUBLIC Boost::log_setup
+  PUBLIC cassandra-cpp-driver::cassandra-cpp-driver
+  PUBLIC fmt::fmt
+  PUBLIC OpenSSL::Crypto
+  PUBLIC OpenSSL::SSL
+  PUBLIC xrpl::libxrpl
+
+  INTERFACE Threads::Threads
+)
+
+if (is_gcc)
+  # FIXME: needed on gcc for now
+  target_compile_definitions (clio PUBLIC BOOST_ASIO_DISABLE_CONCEPTS)
+endif ()
+
+target_sources (clio PRIVATE
   ## Main
   src/main/impl/Build.cpp
   ## Backend
-  src/backend/BackendInterface.cpp
-  src/backend/LedgerCache.cpp
-  ## NextGen Backend
-  src/backend/cassandra/impl/Future.cpp
-  src/backend/cassandra/impl/Cluster.cpp
-  src/backend/cassandra/impl/Batch.cpp
-  src/backend/cassandra/impl/Result.cpp
-  src/backend/cassandra/impl/Tuple.cpp
-  src/backend/cassandra/impl/SslContext.cpp
-  src/backend/cassandra/Handle.cpp
-  src/backend/cassandra/SettingsProvider.cpp
+  src/data/BackendInterface.cpp
+  src/data/LedgerCache.cpp
+  src/data/cassandra/impl/Future.cpp
+  src/data/cassandra/impl/Cluster.cpp
+  src/data/cassandra/impl/Batch.cpp
+  src/data/cassandra/impl/Result.cpp
+  src/data/cassandra/impl/Tuple.cpp
+  src/data/cassandra/impl/SslContext.cpp
+  src/data/cassandra/Handle.cpp
+  src/data/cassandra/SettingsProvider.cpp
   ## ETL
   src/etl/Source.cpp
   src/etl/ProbingSource.cpp
@@ -58,8 +89,10 @@ target_sources(clio PRIVATE
   src/etl/ETLService.cpp
   src/etl/LoadBalancer.cpp
   src/etl/impl/ForwardCache.cpp
-  ## Subscriptions
-  src/subscriptions/SubscriptionManager.cpp
+  ## Feed
+  src/feed/SubscriptionManager.cpp
+  ## Web
+  src/web/IntervalSweepHandler.cpp
   ## RPC
   src/rpc/Errors.cpp
   src/rpc/Factories.cpp
@@ -68,9 +101,10 @@ target_sources(clio PRIVATE
   src/rpc/WorkQueue.cpp
   src/rpc/common/Specs.cpp
   src/rpc/common/Validators.cpp
-  # RPC impl
+  src/rpc/common/MetaProcessors.cpp
+  src/rpc/common/impl/APIVersionParser.cpp
   src/rpc/common/impl/HandlerProvider.cpp
-  ## RPC handler
+  ## RPC handlers
   src/rpc/handlers/AccountChannels.cpp
   src/rpc/handlers/AccountCurrencies.cpp
   src/rpc/handlers/AccountInfo.cpp
@@ -81,6 +115,7 @@ target_sources(clio PRIVATE
   src/rpc/handlers/AccountTx.cpp
   src/rpc/handlers/BookChanges.cpp
   src/rpc/handlers/BookOffers.cpp
+  src/rpc/handlers/DepositAuthorized.cpp
   src/rpc/handlers/GatewayBalances.cpp
   src/rpc/handlers/Ledger.cpp
   src/rpc/handlers/LedgerData.cpp
@@ -96,88 +131,142 @@ target_sources(clio PRIVATE
   src/rpc/handlers/TransactionEntry.cpp
   src/rpc/handlers/Tx.cpp
   ## Util
-  src/config/Config.cpp
-  src/log/Logger.cpp
+  src/util/config/Config.cpp
+  src/util/log/Logger.cpp
   src/util/Taggable.cpp)
 
-add_executable(clio_server src/main/main.cpp)
-target_link_libraries(clio_server PUBLIC clio)
+# Clio server
+add_executable (clio_server src/main/Main.cpp)
+target_link_libraries (clio_server PRIVATE clio)
+target_link_options(clio_server
+  PRIVATE
+    $<$<AND:$<NOT:$<BOOL:${APPLE}>>,$<NOT:$<BOOL:${san}>>>:-static-libstdc++ -static-libgcc>
+)
 
-if(BUILD_TESTS)
-  set(TEST_TARGET clio_tests)
-  add_executable(${TEST_TARGET}
+# Unittesting
+if (tests)
+  set (TEST_TARGET clio_tests)
+  add_executable (${TEST_TARGET}
+    # Common
+    unittests/Main.cpp
     unittests/Playground.cpp
-    unittests/Logger.cpp
-    unittests/Config.cpp
-    unittests/ProfilerTest.cpp
-    unittests/DOSGuard.cpp
-    unittests/SubscriptionTest.cpp
-    unittests/SubscriptionManagerTest.cpp
+    unittests/LoggerTests.cpp
+    unittests/ConfigTests.cpp
+    unittests/ProfilerTests.cpp
+    unittests/JsonUtilTests.cpp
+    unittests/DOSGuardTests.cpp
+    unittests/SubscriptionTests.cpp
+    unittests/SubscriptionManagerTests.cpp
     unittests/util/TestObject.cpp
     unittests/util/StringUtils.cpp
     # ETL
-    unittests/etl/ExtractionDataPipeTest.cpp
-    unittests/etl/ExtractorTest.cpp
-    unittests/etl/TransformerTest.cpp
+    unittests/etl/ExtractionDataPipeTests.cpp
+    unittests/etl/ExtractorTests.cpp
+    unittests/etl/TransformerTests.cpp
+    unittests/etl/CacheLoaderTests.cpp
+    unittests/etl/AmendmentBlockHandlerTests.cpp
     # RPC
     unittests/rpc/ErrorTests.cpp
     unittests/rpc/BaseTests.cpp
-    unittests/rpc/RPCHelpersTest.cpp
-    unittests/rpc/CountersTest.cpp
-    unittests/rpc/AdminVerificationTest.cpp
+    unittests/rpc/RPCHelpersTests.cpp
+    unittests/rpc/CountersTests.cpp
+    unittests/rpc/AdminVerificationTests.cpp
+    unittests/rpc/APIVersionTests.cpp
+    unittests/rpc/ForwardingProxyTests.cpp
+    unittests/rpc/WorkQueueTests.cpp
+    unittests/rpc/AmendmentsTests.cpp
+    unittests/rpc/JsonBoolTests.cpp
     ## RPC handlers
     unittests/rpc/handlers/DefaultProcessorTests.cpp
     unittests/rpc/handlers/TestHandlerTests.cpp
-    unittests/rpc/handlers/AccountCurrenciesTest.cpp
-    unittests/rpc/handlers/AccountLinesTest.cpp
-    unittests/rpc/handlers/AccountTxTest.cpp
-    unittests/rpc/handlers/AccountOffersTest.cpp
-    unittests/rpc/handlers/AccountInfoTest.cpp
-    unittests/rpc/handlers/AccountChannelsTest.cpp
-    unittests/rpc/handlers/AccountNFTsTest.cpp
-    unittests/rpc/handlers/BookOffersTest.cpp
-    unittests/rpc/handlers/GatewayBalancesTest.cpp
-    unittests/rpc/handlers/TxTest.cpp
-    unittests/rpc/handlers/TransactionEntryTest.cpp
-    unittests/rpc/handlers/LedgerEntryTest.cpp
-    unittests/rpc/handlers/LedgerRangeTest.cpp
-    unittests/rpc/handlers/NoRippleCheckTest.cpp
-    unittests/rpc/handlers/ServerInfoTest.cpp
-    unittests/rpc/handlers/PingTest.cpp
-    unittests/rpc/handlers/RandomTest.cpp
-    unittests/rpc/handlers/NFTInfoTest.cpp
-    unittests/rpc/handlers/NFTBuyOffersTest.cpp
-    unittests/rpc/handlers/NFTSellOffersTest.cpp
-    unittests/rpc/handlers/NFTHistoryTest.cpp
-    unittests/rpc/handlers/SubscribeTest.cpp
-    unittests/rpc/handlers/UnsubscribeTest.cpp
-    unittests/rpc/handlers/LedgerDataTest.cpp
-    unittests/rpc/handlers/AccountObjectsTest.cpp
-    unittests/rpc/handlers/BookChangesTest.cpp
-    unittests/rpc/handlers/LedgerTest.cpp
+    unittests/rpc/handlers/AccountCurrenciesTests.cpp
+    unittests/rpc/handlers/AccountLinesTests.cpp
+    unittests/rpc/handlers/AccountTxTests.cpp
+    unittests/rpc/handlers/AccountOffersTests.cpp
+    unittests/rpc/handlers/AccountInfoTests.cpp
+    unittests/rpc/handlers/AccountChannelsTests.cpp
+    unittests/rpc/handlers/AccountNFTsTests.cpp
+    unittests/rpc/handlers/BookOffersTests.cpp
+    unittests/rpc/handlers/DepositAuthorizedTests.cpp
+    unittests/rpc/handlers/GatewayBalancesTests.cpp
+    unittests/rpc/handlers/TxTests.cpp
+    unittests/rpc/handlers/TransactionEntryTests.cpp
+    unittests/rpc/handlers/LedgerEntryTests.cpp
+    unittests/rpc/handlers/LedgerRangeTests.cpp
+    unittests/rpc/handlers/NoRippleCheckTests.cpp
+    unittests/rpc/handlers/ServerInfoTests.cpp
+    unittests/rpc/handlers/PingTests.cpp
+    unittests/rpc/handlers/RandomTests.cpp
+    unittests/rpc/handlers/NFTInfoTests.cpp
+    unittests/rpc/handlers/NFTBuyOffersTests.cpp
+    unittests/rpc/handlers/NFTSellOffersTests.cpp
+    unittests/rpc/handlers/NFTHistoryTests.cpp
+    unittests/rpc/handlers/SubscribeTests.cpp
+    unittests/rpc/handlers/UnsubscribeTests.cpp
+    unittests/rpc/handlers/LedgerDataTests.cpp
+    unittests/rpc/handlers/AccountObjectsTests.cpp
+    unittests/rpc/handlers/BookChangesTests.cpp
+    unittests/rpc/handlers/LedgerTests.cpp
+    unittests/rpc/handlers/VersionHandlerTests.cpp
     # Backend
-    unittests/backend/BackendFactoryTest.cpp
-    unittests/backend/cassandra/BaseTests.cpp
-    unittests/backend/cassandra/BackendTests.cpp
-    unittests/backend/cassandra/RetryPolicyTests.cpp
-    unittests/backend/cassandra/SettingsProviderTests.cpp
-    unittests/backend/cassandra/ExecutionStrategyTests.cpp
-    unittests/backend/cassandra/AsyncExecutorTests.cpp
-    unittests/webserver/ServerTest.cpp
-    unittests/webserver/RPCExecutorTest.cpp)
-  include(CMake/deps/gtest.cmake)
+    unittests/data/BackendFactoryTests.cpp
+    unittests/data/cassandra/BaseTests.cpp
+    unittests/data/cassandra/BackendTests.cpp
+    unittests/data/cassandra/RetryPolicyTests.cpp
+    unittests/data/cassandra/SettingsProviderTests.cpp
+    unittests/data/cassandra/ExecutionStrategyTests.cpp
+    unittests/data/cassandra/AsyncExecutorTests.cpp
+    # Webserver
+    unittests/web/ServerTests.cpp
+    unittests/web/RPCServerHandlerTests.cpp
+    unittests/web/WhitelistHandlerTests.cpp
+    unittests/web/SweepHandlerTests.cpp)
 
-  # test for dwarf5 bug on ci 
-  target_compile_options(clio PUBLIC -gdwarf-4)
+  include (CMake/deps/gtest.cmake)
 
-  # if CODE_COVERAGE enable, add clio_test-ccov
-  if(CODE_COVERAGE)
-    include(CMake/coverage.cmake)
-    add_converage(${TEST_TARGET})
-  endif()
-endif()
+  # See https://github.com/google/googletest/issues/3475
+  gtest_discover_tests (clio_tests DISCOVERY_TIMEOUT 10)
 
-include(CMake/install/install.cmake)
-if(PACKAGING)
-    include(CMake/packaging.cmake)
-endif()
+  # Fix for dwarf5 bug on ci
+  target_compile_options (clio PUBLIC -gdwarf-4)
+
+  target_compile_definitions (${TEST_TARGET} PUBLIC UNITTEST_BUILD)
+  target_include_directories (${TEST_TARGET} PRIVATE unittests)
+  target_link_libraries (${TEST_TARGET} PUBLIC clio gtest::gtest)
+
+  # Generate `clio_tests-ccov` if coverage is enabled
+  # Note: use `make clio_tests-ccov` to generate report
+  if (coverage)
+    target_compile_definitions(${TEST_TARGET} PRIVATE COVERAGE_ENABLED)
+    include (CMake/Coverage.cmake)
+    add_coverage (${TEST_TARGET})
+  endif ()
+endif ()
+
+# Enable selected sanitizer if enabled via `san`
+if (san)
+  target_compile_options (clio
+    PUBLIC
+      # Sanitizers recommend minimum of -O1 for reasonable performance
+      $<$<CONFIG:Debug>:-O1>
+      ${SAN_FLAG}
+      -fno-omit-frame-pointer)
+  target_compile_definitions (clio
+    PUBLIC
+      $<$<STREQUAL:${san},address>:SANITIZER=ASAN>
+      $<$<STREQUAL:${san},thread>:SANITIZER=TSAN>
+      $<$<STREQUAL:${san},memory>:SANITIZER=MSAN>
+      $<$<STREQUAL:${san},undefined>:SANITIZER=UBSAN>)
+  target_link_libraries (clio INTERFACE ${SAN_FLAG} ${SAN_LIB})
+endif ()
+
+# Generate `docs` target for doxygen documentation if enabled
+# Note: use `make docs` to generate the documentation
+if (docs)
+  include (CMake/Docs.cmake)
+endif ()
+
+include (CMake/install/install.cmake)
+if (packaging)
+  include (CMake/packaging.cmake) # This file exists only in build runner
+endif ()

Doxyfile

@@ -1,3 +1,16 @@
 PROJECT_NAME           = "Clio"
-INPUT                  = src
-RECURSIVE              = YES
+INPUT                  = ../src ../unittests
+EXCLUDE_PATTERNS       = *Test*.cpp *Test*.h
+RECURSIVE              = YES
+HAVE_DOT               = YES
+
+QUIET                  = YES
+WARNINGS               = NO
+WARN_NO_PARAMDOC       = NO
+WARN_IF_INCOMPLETE_DOC = NO
+WARN_IF_UNDOCUMENTED   = NO
+
+GENERATE_LATEX         = NO
+GENERATE_HTML          = YES
+
+SORT_MEMBERS_CTORS_1ST = YES

README.md

@@ -1,51 +1,102 @@
 # Clio
-Clio is an XRP Ledger API server. Clio is optimized for RPC calls, over WebSocket or JSON-RPC. Validated
-historical ledger and transaction data are stored in a more space-efficient format,
+
+Clio is an XRP Ledger API server. Clio is optimized for RPC calls, over WebSocket or JSON-RPC. 
+Validated historical ledger and transaction data are stored in a more space-efficient format,
 using up to 4 times less space than rippled. Clio can be configured to store data in Apache Cassandra or ScyllaDB,
-allowing for scalable read throughput. Multiple Clio nodes can share
-access to the same dataset, allowing for a highly available cluster of Clio nodes,
-without the need for redundant data storage or computation.
+allowing for scalable read throughput. Multiple Clio nodes can share access to the same dataset, 
+allowing for a highly available cluster of Clio nodes, without the need for redundant data storage or computation.
 
 Clio offers the full rippled API, with the caveat that Clio by default only returns validated data.
 This means that `ledger_index` defaults to `validated` instead of `current` for all requests.
 Other non-validated data is also not returned, such as information about queued transactions.
-For requests that require access to the p2p network, such as `fee` or `submit`, Clio automatically forwards the request to a rippled node and propagates the response back to the client. To access non-validated data for *any* request, simply add `ledger_index: "current"` to the request, and Clio will forward the request to rippled.
+For requests that require access to the p2p network, such as `fee` or `submit`, Clio automatically forwards the request to a rippled node and propagates the response back to the client. 
+To access non-validated data for *any* request, simply add `ledger_index: "current"` to the request, and Clio will forward the request to rippled.
 
 Clio does not connect to the peer-to-peer network. Instead, Clio extracts data from a group of specified rippled nodes. Running Clio requires access to at least one rippled node
 from which data can be extracted. The rippled node does not need to be running on the same machine as Clio.
 
-
 ## Requirements
 1. Access to a Cassandra cluster or ScyllaDB cluster. Can be local or remote.
-
 2. Access to one or more rippled nodes. Can be local or remote.
 
 ## Building
 
-Clio is built with CMake. Clio requires at least GCC-11/clang-14.0.0 (C++20), and Boost 1.75.0.
+Clio is built with CMake and uses Conan for managing dependencies. 
+It is written in C++20 and therefore requires a modern compiler.
 
-Use these instructions to build a Clio executable from the source. These instructions were tested on Ubuntu 20.04 LTS.
+## Prerequisites 
 
-```sh
-# Install dependencies
-  sudo apt-get -y install git pkg-config protobuf-compiler libprotobuf-dev libssl-dev wget build-essential bison flex autoconf cmake clang-format
-# Install gcovr to run code coverage
-  sudo apt-get -y install gcovr
+### Minimum Requirements
 
-# Compile Boost
-  wget -O $HOME/boost_1_75_0.tar.gz https://boostorg.jfrog.io/artifactory/main/release/1.75.0/source/boost_1_75_0.tar.gz
-  tar xvzf $HOME/boost_1_75_0.tar.gz
-  cd $HOME/boost_1_75_0
-  ./bootstrap.sh
-  ./b2 -j$(nproc)
-  echo "export BOOST_ROOT=$HOME/boost_1_75_0" >> $HOME/.profile && source $HOME/.profile
+- [Python 3.7](https://www.python.org/downloads/)
+- [Conan 1.55](https://conan.io/downloads.html)
+- [CMake 3.16](https://cmake.org/download/)
+- [**Optional**] [GCovr](https://gcc.gnu.org/onlinedocs/gcc/Gcov.html) (needed for code coverage generation)
 
-# Clone the Clio Git repository & build Clio
-  cd $HOME
-  git clone https://github.com/XRPLF/clio.git
-  cd $HOME/clio
-  cmake -B build && cmake --build build --parallel $(nproc)
+| Compiler    | Version |
+|-------------|---------|
+| GCC         | 11      |
+| Clang       | 14      |
+| Apple Clang | 14.0.3  |
+
+### Conan configuration
+
+Clio does not require anything but default settings in your (`~/.conan/profiles/default`) Conan profile. It's best to have no extra flags specified. 
+> Mac example:
 ```
+[settings]
+os=Macos
+os_build=Macos
+arch=armv8
+arch_build=armv8
+compiler=apple-clang
+compiler.version=14
+compiler.libcxx=libc++
+build_type=Release
+compiler.cppstd=20
+```
+> Linux example:
+```
+[settings]
+os=Linux
+os_build=Linux
+arch=x86_64
+arch_build=x86_64
+compiler=gcc
+compiler.version=11
+compiler.libcxx=libstdc++11
+build_type=Release
+compiler.cppstd=20
+```
+
+### Artifactory
+
+1. Make sure artifactory is setup with Conan
+```sh
+conan remote add --insert 0 conan-non-prod http://18.143.149.228:8081/artifactory/api/conan/conan-non-prod
+```
+Now you should be able to download prebuilt `xrpl` package on some platforms.
+
+2. Remove old packages you may have cached: 
+```sh 
+conan remove -f xrpl
+conan remove -f cassandra-cpp-driver
+```
+
+## Building Clio
+
+Navigate to Clio's root directory and perform
+```sh
+mkdir build && cd build
+conan install .. --output-folder . --build missing --settings build_type=Release -o tests=True
+cmake -DCMAKE_TOOLCHAIN_FILE:FILEPATH=build/generators/conan_toolchain.cmake -DCMAKE_BUILD_TYPE=Release ..
+cmake --build . --parallel 8 # or without the number if you feel extra adventurous
+```
+If all goes well, `conan install` will find required packages and `cmake` will do the rest. you should end up with `clio_server` and `clio_tests` in the `build` directory (the current directory).
+
+> **Tip:** You can omit the `-o tests=True` in `conan install` command above if you don't want to build `clio_tests`.
+
+> **Tip:** To generate a Code Coverage report, include `-o coverage=True` in the `conan install` command above, along with `-o tests=True` to enable tests. After running the `cmake` commands, execute `make clio_tests-ccov`. The coverage report will be found at `clio_tests-llvm-cov/index.html`.
 
 ## Running
 ```sh
@@ -96,12 +147,12 @@ The parameters `ssl_cert_file` and `ssl_key_file` can also be added to the top l
 An example of how to specify `ssl_cert_file` and `ssl_key_file` in the config:
 
 ```json
-"server":{
+"server": {
     "ip": "0.0.0.0",
     "port": 51233
 },
-"ssl_cert_file" : "/full/path/to/cert.file",
-"ssl_key_file" : "/full/path/to/key.file"
+"ssl_cert_file": "/full/path/to/cert.file",
+"ssl_key_file": "/full/path/to/key.file"
 ```
 
 Once your config files are ready, start rippled and Clio. It doesn't matter which you
@@ -172,6 +223,20 @@ which can cause high latencies. A possible alternative to this is to just deploy
 a database in each region, and the Clio nodes in each region use their region's database.
 This is effectively two systems.
 
+Clio supports API versioning as [described here](https://xrpl.org/request-formatting.html#api-versioning).
+It's possible to configure `minimum`, `maximum` and `default` version like so:
+```json
+"api_version": {
+    "min": 1,
+    "max": 2,
+    "default": 1
+}
+```
+All of the above are optional.
+Clio will fallback to hardcoded defaults when not specified in the config file or configured values are outside
+of the minimum and maximum supported versions hardcoded in `src/rpc/common/APIVersion.h`.
+> **Note:** See `example-config.json` for more details. 
+
 ## Developing against `rippled` in standalone mode
 
 If you wish you develop against a `rippled` instance running in standalone

cloud-example-config.json

@@ -1,14 +1,22 @@
+/*
+ * This is an example configuration file. Please do not use without modifying to suit your needs.
+ */
 {
     "database": {
         "type": "cassandra",
         "cassandra": {
+            // This option can be used to setup a secure connect bundle connection
             "secure_connect_bundle": "[path/to/zip. ignore if using contact_points]",
+            // The following options are used only if using contact_points
             "contact_points": "[ip. ignore if using secure_connect_bundle]",
             "port": "[port. ignore if using_secure_connect_bundle]",
-            "keyspace": "clio",
+            // Authentication settings
             "username": "[username, if any]",
             "password": "[password, if any]",
-            "max_requests_outstanding": 25000,
+            // Other common settings
+            "keyspace": "clio",
+            "max_write_requests_outstanding": 25000,
+            "max_read_requests_outstanding": 30000,
             "threads": 8
         }
     },

conanfile.py

@@ -0,0 +1,87 @@
+from conan import ConanFile
+from conan.tools.cmake import CMake, CMakeToolchain, cmake_layout
+import re
+
+class Clio(ConanFile):
+    name = 'clio'
+    license = 'ISC'
+    author = 'Alex Kremer <akremer@ripple.com>, John Freeman <jfreeman@ripple.com>'
+    url = 'https://github.com/xrplf/clio'
+    description = 'Clio RPC server'
+    settings = 'os', 'compiler', 'build_type', 'arch'
+    options = {
+        'fPIC': [True, False],
+        'verbose': [True, False],   
+        'tests': [True, False],     # build unit tests; create `clio_tests` binary
+        'docs': [True, False],      # doxygen API docs; create custom target 'docs'
+        'packaging': [True, False], # create distribution packages
+        'coverage': [True, False],  # build for test coverage report; create custom target `clio_tests-ccov`
+    }
+
+    requires = [
+        'boost/1.82.0',
+        'cassandra-cpp-driver/2.16.2',
+        'fmt/10.0.0',
+        'grpc/1.50.1',
+        'openssl/1.1.1u',
+        'xrpl/1.12.0',
+    ]
+
+    default_options = {
+        'fPIC': True,
+        'verbose': False,
+        'tests': False,
+        'packaging': False,
+        'coverage': False,
+        'docs': False,
+        
+        'xrpl/*:tests': False,
+        'cassandra-cpp-driver/*:shared': False,
+        'date/*:header_only': True,
+        'grpc/*:shared': False,
+        'grpc/*:secure': True,
+        'libpq/*:shared': False,
+        'lz4/*:shared': False,
+        'openssl/*:shared': False,
+        'protobuf/*:shared': False,
+        'protobuf/*:with_zlib': True,
+        'snappy/*:shared': False,
+        'gtest/*:no_main': True,
+    }
+
+    exports_sources = (
+        'CMakeLists.txt', 'CMake/*', 'src/*'
+    )
+
+    def requirements(self):
+        if self.options.tests:
+            self.requires('gtest/1.13.0')
+
+    def configure(self):
+        if self.settings.compiler == 'apple-clang':
+            self.options['boost'].visibility = 'global'
+
+    def layout(self):
+        cmake_layout(self)
+        # Fix this setting to follow the default introduced in Conan 1.48 
+        # to align with our build instructions.
+        self.folders.generators = 'build/generators'
+
+    generators = 'CMakeDeps'
+    def generate(self):
+        tc = CMakeToolchain(self)
+        tc.variables['verbose'] = self.options.verbose
+        tc.variables['tests'] = self.options.tests
+        tc.variables['coverage'] = self.options.coverage
+        tc.variables['docs'] = self.options.docs
+        tc.variables['packaging'] = self.options.packaging
+        tc.generate()
+
+    def build(self):
+        cmake = CMake(self)
+        cmake.configure()
+        cmake.build()
+
+    def package(self):
+        cmake = CMake(self)
+        cmake.install()

example-config.json

@@ -1,3 +1,6 @@
+/*
+ * This is an example configuration file. Please do not use without modifying to suit your needs.
+ */
 {
     "database": {
         "type": "cassandra",
@@ -9,7 +12,18 @@
             "table_prefix": "",
             "max_write_requests_outstanding": 25000,
             "max_read_requests_outstanding": 30000,
-            "threads": 8
+            "threads": 8,
+            //
+            // Advanced options. USE AT OWN RISK:
+            // ---
+            "core_connections_per_host": 1 // Defaults to 1
+            //
+            // Below options will use defaults from cassandra driver if left unspecified.
+            // See https://docs.datastax.com/en/developer/cpp-driver/2.17/api/struct.CassCluster/ for details.
+            // 
+            // "queue_size_io": 2
+            //
+            // ---
         }
     },
     "etl_sources": [
@@ -20,21 +34,24 @@
         }
     ],
     "dos_guard": {
+        // Comma-separated list of IPs to exclude from rate limiting
         "whitelist": [
             "127.0.0.1"
-        ], // comma-separated list of ips to exclude from rate limiting
-        /* The below values are the default values and are only specified here
-         * for documentation purposes. The rate limiter currently limits
-         * connections and bandwidth per ip. The rate limiter looks at the raw
-         * ip of a client connection, and so requests routed through a load
-         * balancer will all have the same ip and be treated as a single client
-         */
-        "max_fetches": 1000000, // max bytes per ip per sweep interval
-        "max_connections": 20, // max connections per ip
-        "max_requests": 20, // max connections per ip
-        "sweep_interval": 1 // time in seconds before resetting bytes per ip count
+        ],
+        //
+        // The below values are the default values and are only specified here
+        // for documentation purposes. The rate limiter currently limits
+        // connections and bandwidth per IP. The rate limiter looks at the raw
+        // IP of a client connection, and so requests routed through a load
+        // balancer will all have the same IP and be treated as a single client.
+        //
+        "max_fetches": 1000000, // Max bytes per IP per sweep interval
+        "max_connections": 20, // Max connections per IP
+        "max_requests": 20, // Max connections per IP per sweep interval
+        "sweep_interval": 1 // Time in seconds before resetting max_fetches and max_requests
     },
     "cache": {
+        // Comma-separated list of peer nodes that Clio can use to download cache from at startup
         "peers": [
             {
                 "ip": "127.0.0.1",
@@ -45,11 +62,12 @@
     "server": {
         "ip": "0.0.0.0",
         "port": 51233,
-        /* Max number of requests to queue up before rejecting further requests.
-        * Defaults to 0, which disables the limit
-        */
+        // Max number of requests to queue up before rejecting further requests.
+        // Defaults to 0, which disables the limit.
         "max_queue_size": 500
     },
+    // Overrides log level on a per logging channel.
+    // Defaults to global "log_level" for each unspecified channel.
     "log_channels": [
         {
             "channel": "Backend",
@@ -77,17 +95,24 @@
         }
     ],
     "log_level": "info",
-    "log_format": "%TimeStamp% (%SourceLocation%) [%ThreadID%] %Channel%:%Severity% %Message%", // This is the default format
+    // Log format (this is the default format)
+    "log_format": "%TimeStamp% (%SourceLocation%) [%ThreadID%] %Channel%:%Severity% %Message%",
     "log_to_console": true,
-    "log_directory": "./clio_log",
+    // Clio logs to file in the specified directory only if "log_directory" is set
+    // "log_directory": "./clio_log",
     "log_rotation_size": 2048,
     "log_directory_max_size": 51200,
     "log_rotation_hour_interval": 12,
     "log_tag_style": "uint",
     "extractor_threads": 8,
     "read_only": false,
-    //"start_sequence": [integer] the ledger index to start from,
-    //"finish_sequence": [integer] the ledger index to finish at,
-    //"ssl_cert_file" : "/full/path/to/cert.file",
-    //"ssl_key_file" : "/full/path/to/key.file"
+    // "start_sequence": [integer] the ledger index to start from,
+    // "finish_sequence": [integer] the ledger index to finish at,
+    // "ssl_cert_file" : "/full/path/to/cert.file",
+    // "ssl_key_file" : "/full/path/to/key.file"
+    "api_version": {
+        "min": 1, // Minimum API version supported (could be 1 or 2)
+        "max": 2, // Maximum API version supported (could be 1 or 2, but >= min)
+        "default": 1 // Clio behaves the same as rippled by default
+    }
 }

external/cassandra/conanfile.py

@@ -0,0 +1,87 @@
+from conan import ConanFile, tools
+from conan.tools.cmake import CMake, CMakeToolchain
+
+class Cassandra(ConanFile):
+    name = 'cassandra-cpp-driver'
+    version = '2.16.2'
+    license = 'Apache-2.0'
+    url = 'https://github.com/conan-io/conan-center-index'
+    homepage = 'https://docs.datastax.com/en/developer/cpp-driver/'
+    description = 'Cassandra C++ Driver'
+    topics = ('conan', 'cassandra', 'driver')
+
+    settings = 'os', 'arch', 'compiler', 'build_type'
+    options = {
+        'shared': [True, False],
+        'fPIC': [True, False],
+        'install_header_in_subdir': [True, False],
+        'use_atomic': [None, 'boost', 'std'],
+        'with_openssl': [True, False],
+        'with_zlib': [True, False],
+        'with_kerberos': [True, False],
+        'use_timerfd': [True, False],
+    }
+    default_options = {
+        'shared': False,
+        'fPIC': True,
+        'install_header_in_subdir': False,
+        'use_atomic': None,
+        'with_openssl': True,
+        'with_zlib': True,
+        'with_kerberos': False,
+        'use_timerfd': True,
+    }
+
+    def requirements(self):
+        self.requires('libuv/1.44.1')
+        self.requires('http_parser/2.9.4')
+        if self.options.with_openssl:
+            self.requires('openssl/1.1.1q')
+        if self.options.with_zlib:
+            self.requires('minizip/1.2.12')
+            self.requires('zlib/1.2.13')
+        if self.options.use_atomic == 'boost':
+            self.requires('boost/1.79.0')
+
+    exports_sources = ['CMakeLists.txt']
+
+    def config_options(self):
+        if self.settings.os == 'Windows':
+            del self.options.fPIC
+
+    def configure(self):
+        self.options['libuv'].shared = self.options.shared
+
+    def generate(self):
+        tc = CMakeToolchain(self)
+        if self.settings.get_safe('compiler.cppstd') == '20':
+            tc.blocks['cppstd'].values['cppstd'] = 17
+        tc.variables['CASS_BUILD_STATIC'] = not self.options.shared
+        tc.variables['CASS_USE_STATIC_LIBS'] = not self.options.shared
+        tc.variables['CASS_BUILD_SHARED'] = self.options.shared
+        tc.variables['LIBUV_ROOT_DIR'] = self.deps_cpp_info['libuv'].rootpath
+        if self.options.with_openssl:
+            tc.variables['OPENSSL_ROOT_DIR'] = self.deps_cpp_info['openssl'].rootpath
+        tc.generate()
+
+    def source(self):
+        tools.files.get(self, f'https://github.com/datastax/cpp-driver/archive/refs/tags/{self.version}.tar.gz', strip_root=True)
+
+    def build(self):
+        cmake = CMake(self)
+        cmake.configure()
+        cmake.build()
+
+    def package(self):
+        cmake = CMake(self)
+        cmake.install()
+
+    def package_info(self):
+        if self.options.shared:
+            self.cpp_info.libs = ['cassandra']
+        else:
+            self.cpp_info.libs = ['cassandra_static']
+        if self.settings.os == 'Windows':
+            self.cpp_info.libs.extend(['iphlpapi', 'psapi', 'wsock32', 'crypt32', 'ws2_32', 'userenv'])
+            if not self.options.shared:
+                self.cpp_info.defines = ['CASS_STATIC']

src/backend/BackendInterface.h

@@ -1,583 +0,0 @@
-//------------------------------------------------------------------------------
-/*
-    This file is part of clio: https://github.com/XRPLF/clio
-    Copyright (c) 2022, the clio developers.
-
-    Permission to use, copy, modify, and distribute this software for any
-    purpose with or without fee is hereby granted, provided that the above
-    copyright notice and this permission notice appear in all copies.
-
-    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
-    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
-    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
-    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
-    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
-    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
-    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
-*/
-//==============================================================================
-
-#pragma once
-
-#include <ripple/ledger/ReadView.h>
-#include <backend/DBHelpers.h>
-#include <backend/LedgerCache.h>
-#include <backend/Types.h>
-#include <config/Config.h>
-#include <log/Logger.h>
-
-#include <boost/asio/spawn.hpp>
-#include <boost/json.hpp>
-
-#include <thread>
-#include <type_traits>
-
-namespace Backend {
-
-/**
- * @brief Throws an error when database read time limit is exceeded.
- *
- * This class is throws an error when read time limit is exceeded but
- * is also paired with a separate class to retry the connection.
- */
-class DatabaseTimeout : public std::exception
-{
-public:
-    const char*
-    what() const throw() override
-    {
-        return "Database read timed out. Please retry the request";
-    }
-};
-
-/**
- * @brief Separate class that reattempts connection after time limit.
- *
- * @tparam F Represents a class of handlers for Cassandra database.
- * @param func Instance of Cassandra database handler class.
- * @param waitMs Is the arbitrary time limit of 500ms.
- * @return auto
- */
-template <class F>
-auto
-retryOnTimeout(F func, size_t waitMs = 500)
-{
-    static clio::Logger log{"Backend"};
-
-    while (true)
-    {
-        try
-        {
-            return func();
-        }
-        catch (DatabaseTimeout& t)
-        {
-            log.error() << "Database request timed out. Sleeping and retrying ... ";
-            std::this_thread::sleep_for(std::chrono::milliseconds(waitMs));
-        }
-    }
-}
-
-/**
- * @brief Passes in serialized handlers in an asynchronous fashion.
- *
- * Note that the synchronous auto passes handlers critical to supporting
- * the Clio backend. The coroutine types are checked if same/different.
- *
- * @tparam F Represents a class of handlers for Cassandra database.
- * @param f R-value instance of Cassandra handler class.
- * @return auto
- */
-template <class F>
-auto
-synchronous(F&& f)
-{
-    /** @brief Serialized handlers and their execution.
-     *
-     * The ctx class is converted into a serialized handler, also named
-     * ctx, and is used to pass a stream of data into the method.
-     */
-    boost::asio::io_context ctx;
-    boost::asio::io_context::strand strand(ctx);
-    std::optional<boost::asio::io_context::work> work;
-
-    /*! @brief Place the ctx within the vector of serialized handlers. */
-    work.emplace(ctx);
-
-    /**
-     * @brief If/else statements regarding coroutine type matching.
-     *
-     * R is the currently executing coroutine that is about to get passed.
-     * If corountine types do not match, the current one's type is stored.
-     */
-    using R = typename boost::result_of<F(boost::asio::yield_context&)>::type;
-    if constexpr (!std::is_same<R, void>::value)
-    {
-        /**
-         * @brief When the coroutine type is the same
-         *
-         * The spawn function enables programs to implement asynchronous logic
-         * in a synchronous manner. res stores the instance of the currently
-         * executing coroutine, yield. The different type is returned.
-         */
-        R res;
-        boost::asio::spawn(strand, [&f, &work, &res](boost::asio::yield_context yield) {
-            res = f(yield);
-            work.reset();
-        });
-
-        ctx.run();
-        return res;
-    }
-    else
-    {
-        /*! @brief When the corutine type is different, run as normal. */
-        boost::asio::spawn(strand, [&f, &work](boost::asio::yield_context yield) {
-            f(yield);
-            work.reset();
-        });
-
-        ctx.run();
-    }
-}
-
-/**
- * @brief Reestablishes synchronous connection on timeout.
- *
- * @tparam Represents a class of handlers for Cassandra database.
- * @param f R-value instance of Cassandra database handler class.
- * @return auto
- */
-template <class F>
-auto
-synchronousAndRetryOnTimeout(F&& f)
-{
-    return retryOnTimeout([&]() { return synchronous(f); });
-}
-
-/*! @brief Handles ledger and transaction backend data. */
-class BackendInterface
-{
-    /**
-     * @brief Shared mutexes and a cache for the interface.
-     *
-     * rngMutex is a shared mutex. Shared mutexes prevent shared data
-     * from being accessed by multiple threads and has two levels of
-     * access: shared and exclusive.
-     */
-protected:
-    mutable std::shared_mutex rngMtx_;
-    std::optional<LedgerRange> range;
-    LedgerCache cache_;
-
-    /**
-     * @brief Public read methods
-     *
-     * All of these reads methods can throw DatabaseTimeout. When writing
-     * code in an RPC handler, this exception does not need to be caught:
-     * when an RPC results in a timeout, an error is returned to the client.
-     */
-
-public:
-    BackendInterface() = default;
-    virtual ~BackendInterface() = default;
-
-    /**
-     * @brief Cache that holds states of the ledger
-     * @return Immutable cache
-     */
-    LedgerCache const&
-    cache() const
-    {
-        return cache_;
-    }
-
-    /**
-     * @brief Cache that holds states of the ledger
-     * @return Mutable cache
-     */
-    LedgerCache&
-    cache()
-    {
-        return cache_;
-    }
-
-    /*! @brief Fetches a specific ledger by sequence number. */
-    virtual std::optional<ripple::LedgerInfo>
-    fetchLedgerBySequence(std::uint32_t const sequence, boost::asio::yield_context& yield) const = 0;
-
-    /*! @brief Fetches a specific ledger by hash. */
-    virtual std::optional<ripple::LedgerInfo>
-    fetchLedgerByHash(ripple::uint256 const& hash, boost::asio::yield_context& yield) const = 0;
-
-    /*! @brief Fetches the latest ledger sequence. */
-    virtual std::optional<std::uint32_t>
-    fetchLatestLedgerSequence(boost::asio::yield_context& yield) const = 0;
-
-    /*! @brief Fetches the current ledger range while locking that process */
-    std::optional<LedgerRange>
-    fetchLedgerRange() const
-    {
-        std::shared_lock lck(rngMtx_);
-        return range;
-    }
-
-    /**
-     * @brief Updates the range of sequences to be tracked.
-     *
-     * Function that continues updating the range sliding window or creates
-     * a new sliding window once the maxSequence limit has been reached.
-     *
-     * @param newMax Unsigned 32-bit integer representing new max of range.
-     */
-    void
-    updateRange(uint32_t newMax)
-    {
-        std::scoped_lock lck(rngMtx_);
-        assert(!range || newMax >= range->maxSequence);
-        if (!range)
-            range = {newMax, newMax};
-        else
-            range->maxSequence = newMax;
-    }
-
-    /**
-     * @brief Returns the fees for specific transactions.
-     *
-     * @param seq Unsigned 32-bit integer reprsenting sequence.
-     * @param yield The currently executing coroutine.
-     * @return std::optional<ripple::Fees>
-     */
-    std::optional<ripple::Fees>
-    fetchFees(std::uint32_t const seq, boost::asio::yield_context& yield) const;
-
-    /*! @brief TRANSACTION METHODS */
-    /**
-     * @brief Fetches a specific transaction.
-     *
-     * @param hash Unsigned 256-bit integer representing hash.
-     * @param yield The currently executing coroutine.
-     * @return std::optional<TransactionAndMetadata>
-     */
-    virtual std::optional<TransactionAndMetadata>
-    fetchTransaction(ripple::uint256 const& hash, boost::asio::yield_context& yield) const = 0;
-
-    /**
-     * @brief Fetches multiple transactions.
-     *
-     * @param hashes Unsigned integer value representing a hash.
-     * @param yield The currently executing coroutine.
-     * @return std::vector<TransactionAndMetadata>
-     */
-    virtual std::vector<TransactionAndMetadata>
-    fetchTransactions(std::vector<ripple::uint256> const& hashes, boost::asio::yield_context& yield) const = 0;
-
-    /**
-     * @brief Fetches all transactions for a specific account
-     *
-     * @param account A specific XRPL Account, speciifed by unique type
-     * accountID.
-     * @param limit Paging limit for how many transactions can be returned per
-     * page.
-     * @param forward Boolean whether paging happens forwards or backwards.
-     * @param cursor Important metadata returned every time paging occurs.
-     * @param yield Currently executing coroutine.
-     * @return TransactionsAndCursor
-     */
-    virtual TransactionsAndCursor
-    fetchAccountTransactions(
-        ripple::AccountID const& account,
-        std::uint32_t const limit,
-        bool forward,
-        std::optional<TransactionsCursor> const& cursor,
-        boost::asio::yield_context& yield) const = 0;
-
-    /**
-     * @brief Fetches all transactions from a specific ledger.
-     *
-     * @param ledgerSequence Unsigned 32-bit integer for latest total
-     * transactions.
-     * @param yield Currently executing coroutine.
-     * @return std::vector<TransactionAndMetadata>
-     */
-    virtual std::vector<TransactionAndMetadata>
-    fetchAllTransactionsInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context& yield) const = 0;
-
-    /**
-     * @brief Fetches all transaction hashes from a specific ledger.
-     *
-     * @param ledgerSequence Standard unsigned integer.
-     * @param yield Currently executing coroutine.
-     * @return std::vector<ripple::uint256>
-     */
-    virtual std::vector<ripple::uint256>
-    fetchAllTransactionHashesInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context& yield) const = 0;
-
-    /*! @brief NFT methods */
-    /**
-     * @brief Fetches a specific NFT
-     *
-     * @param tokenID Unsigned 256-bit integer.
-     * @param ledgerSequence Standard unsigned integer.
-     * @param yield Currently executing coroutine.
-     * @return std::optional<NFT>
-     */
-    virtual std::optional<NFT>
-    fetchNFT(ripple::uint256 const& tokenID, std::uint32_t const ledgerSequence, boost::asio::yield_context& yield)
-        const = 0;
-
-    /**
-     * @brief Fetches all transactions for a specific NFT.
-     *
-     * @param tokenID Unsigned 256-bit integer.
-     * @param limit Paging limit as to how many transactions return per page.
-     * @param forward Boolean whether paging happens forwards or backwards.
-     * @param cursorIn Represents transaction number and ledger sequence.
-     * @param yield Currently executing coroutine is passed in as input.
-     * @return TransactionsAndCursor
-     */
-    virtual TransactionsAndCursor
-    fetchNFTTransactions(
-        ripple::uint256 const& tokenID,
-        std::uint32_t const limit,
-        bool const forward,
-        std::optional<TransactionsCursor> const& cursorIn,
-        boost::asio::yield_context& yield) const = 0;
-
-    /*! @brief STATE DATA METHODS */
-    /**
-     * @brief Fetches a specific ledger object: vector of unsigned chars
-     *
-     * @param key Unsigned 256-bit integer.
-     * @param sequence Unsigned 32-bit integer.
-     * @param yield Currently executing coroutine.
-     * @return std::optional<Blob>
-     */
-    std::optional<Blob>
-    fetchLedgerObject(ripple::uint256 const& key, std::uint32_t const sequence, boost::asio::yield_context& yield)
-        const;
-
-    /**
-     * @brief Fetches all ledger objects: a vector of vectors of unsigned chars.
-     *
-     * @param keys Unsigned 256-bit integer.
-     * @param sequence Unsigned 32-bit integer.
-     * @param yield Currently executing coroutine.
-     * @return std::vector<Blob>
-     */
-    std::vector<Blob>
-    fetchLedgerObjects(
-        std::vector<ripple::uint256> const& keys,
-        std::uint32_t const sequence,
-        boost::asio::yield_context& yield) const;
-
-    /*! @brief Virtual function version of fetchLedgerObject */
-    virtual std::optional<Blob>
-    doFetchLedgerObject(ripple::uint256 const& key, std::uint32_t const sequence, boost::asio::yield_context& yield)
-        const = 0;
-
-    /*! @brief Virtual function version of fetchLedgerObjects */
-    virtual std::vector<Blob>
-    doFetchLedgerObjects(
-        std::vector<ripple::uint256> const& keys,
-        std::uint32_t const sequence,
-        boost::asio::yield_context& yield) const = 0;
-
-    /**
-     * @brief Returns the difference between ledgers: vector of objects
-     *
-     * Objects are made of a key value, vector of unsigned chars (blob),
-     * and a boolean detailing whether keys and blob match.
-     *
-     * @param ledgerSequence Standard unsigned integer.
-     * @param yield Currently executing coroutine.
-     * @return std::vector<LedgerObject>
-     */
-    virtual std::vector<LedgerObject>
-    fetchLedgerDiff(std::uint32_t const ledgerSequence, boost::asio::yield_context& yield) const = 0;
-
-    /**
-     * @brief Fetches a page of ledger objects, ordered by key/index.
-     *
-     * @param cursor Important metadata returned every time paging occurs.
-     * @param ledgerSequence Standard unsigned integer.
-     * @param limit Paging limit as to how many transactions returned per page.
-     * @param outOfOrder Boolean on whether ledger page is out of order.
-     * @param yield Currently executing coroutine.
-     * @return LedgerPage
-     */
-    LedgerPage
-    fetchLedgerPage(
-        std::optional<ripple::uint256> const& cursor,
-        std::uint32_t const ledgerSequence,
-        std::uint32_t const limit,
-        bool outOfOrder,
-        boost::asio::yield_context& yield) const;
-
-    /*! @brief Fetches successor object from key/index. */
-    std::optional<LedgerObject>
-    fetchSuccessorObject(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context& yield)
-        const;
-
-    /*! @brief Fetches successor key from key/index. */
-    std::optional<ripple::uint256>
-    fetchSuccessorKey(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context& yield) const;
-
-    /*! @brief Virtual function version of fetchSuccessorKey. */
-    virtual std::optional<ripple::uint256>
-    doFetchSuccessorKey(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context& yield)
-        const = 0;
-
-    /**
-     * @brief Fetches book offers.
-     *
-     * @param book Unsigned 256-bit integer.
-     * @param ledgerSequence Standard unsigned integer.
-     * @param limit Pagaing limit as to how many transactions returned per page.
-     * @param cursor Important metadata returned every time paging occurs.
-     * @param yield Currently executing coroutine.
-     * @return BookOffersPage
-     */
-    BookOffersPage
-    fetchBookOffers(
-        ripple::uint256 const& book,
-        std::uint32_t const ledgerSequence,
-        std::uint32_t const limit,
-        boost::asio::yield_context& yield) const;
-
-    /**
-     * @brief Returns a ledger range
-     *
-     * Ledger range is a struct of min and max sequence numbers). Due to
-     * the use of [&], which denotes a special case of a lambda expression
-     * where values found outside the scope are passed by reference, wrt the
-     * currently executing coroutine.
-     *
-     * @return std::optional<LedgerRange>
-     */
-    std::optional<LedgerRange>
-    hardFetchLedgerRange() const
-    {
-        return synchronous([&](boost::asio::yield_context yield) { return hardFetchLedgerRange(yield); });
-    }
-
-    /*! @brief Virtual function equivalent of hardFetchLedgerRange. */
-    virtual std::optional<LedgerRange>
-    hardFetchLedgerRange(boost::asio::yield_context& yield) const = 0;
-
-    /*! @brief Fetches ledger range but doesn't throw timeout. Use with care. */
-    std::optional<LedgerRange>
-    hardFetchLedgerRangeNoThrow() const;
-    /*! @brief Fetches ledger range but doesn't throw timeout. Use with care. */
-    std::optional<LedgerRange>
-    hardFetchLedgerRangeNoThrow(boost::asio::yield_context& yield) const;
-
-    /**
-     * @brief Writes to a specific ledger.
-     *
-     * @param ledgerInfo Const on ledger information.
-     * @param ledgerHeader r-value string representing ledger header.
-     */
-    virtual void
-    writeLedger(ripple::LedgerInfo const& ledgerInfo, std::string&& ledgerHeader) = 0;
-
-    /**
-     * @brief Writes a new ledger object.
-     *
-     * The key and blob are r-value references and do NOT have memory addresses.
-     *
-     * @param key String represented as an r-value.
-     * @param seq Unsigned integer representing a sequence.
-     * @param blob r-value vector of unsigned characters (blob).
-     */
-    virtual void
-    writeLedgerObject(std::string&& key, std::uint32_t const seq, std::string&& blob);
-
-    /**
-     * @brief Writes a new transaction.
-     *
-     * @param hash r-value reference. No memory address.
-     * @param seq Unsigned 32-bit integer.
-     * @param date Unsigned 32-bit integer.
-     * @param transaction r-value reference. No memory address.
-     * @param metadata r-value refrence. No memory address.
-     */
-    virtual void
-    writeTransaction(
-        std::string&& hash,
-        std::uint32_t const seq,
-        std::uint32_t const date,
-        std::string&& transaction,
-        std::string&& metadata) = 0;
-
-    /**
-     * @brief Write a new NFT.
-     *
-     * @param data Passed in as an r-value reference.
-     */
-    virtual void
-    writeNFTs(std::vector<NFTsData>&& data) = 0;
-
-    /**
-     * @brief Write a new set of account transactions.
-     *
-     * @param data Passed in as an r-value reference.
-     */
-    virtual void
-    writeAccountTransactions(std::vector<AccountTransactionsData>&& data) = 0;
-
-    /**
-     * @brief Write a new transaction for a specific NFT.
-     *
-     * @param data Passed in as an r-value reference.
-     */
-    virtual void
-    writeNFTTransactions(std::vector<NFTTransactionsData>&& data) = 0;
-
-    /**
-     * @brief Write a new successor.
-     *
-     * @param key Passed in as an r-value reference.
-     * @param seq Unsigned 32-bit integer.
-     * @param successor Passed in as an r-value reference.
-     */
-    virtual void
-    writeSuccessor(std::string&& key, std::uint32_t const seq, std::string&& successor) = 0;
-
-    /*! @brief Tells database we will write data for a specific ledger. */
-    virtual void
-    startWrites() const = 0;
-
-    /**
-     * @brief Tells database we finished writing all data for a specific ledger.
-     *
-     * TODO: change the return value to represent different results:
-     * Committed, write conflict, errored, successful but not committed
-     *
-     * @param ledgerSequence Const unsigned 32-bit integer on ledger sequence.
-     * @return true
-     * @return false
-     */
-    bool
-    finishWrites(std::uint32_t const ledgerSequence);
-
-    virtual bool
-    isTooBusy() const = 0;
-
-private:
-    /**
-     * @brief Private helper method to write ledger object
-     *
-     * @param key r-value string representing key.
-     * @param seq Unsigned 32-bit integer representing sequence.
-     * @param blob r-value vector of unsigned chars.
-     */
-    virtual void
-    doWriteLedgerObject(std::string&& key, std::uint32_t const seq, std::string&& blob) = 0;
-
-    virtual bool
-    doFinishWrites() = 0;
-};
-
-}  // namespace Backend
-using BackendInterface = Backend::BackendInterface;

src/data/BackendFactory.h

@@ -19,19 +19,26 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
-#include <backend/CassandraBackend.h>
-#include <config/Config.h>
-#include <log/Logger.h>
+#include <data/BackendInterface.h>
+#include <data/CassandraBackend.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
 
 #include <boost/algorithm/string.hpp>
 
-namespace Backend {
+namespace data {
+
+/**
+ * @brief A factory function that creates the backend based on a config.
+ *
+ * @param config The clio config to use
+ * @return A shared_ptr<BackendInterface> with the selected implementation
+ */
 std::shared_ptr<BackendInterface>
-make_Backend(boost::asio::io_context& ioc, clio::Config const& config)
+make_Backend(util::Config const& config)
 {
-    static clio::Logger log{"Backend"};
-    log.info() << "Constructing BackendInterface";
+    static util::Logger log{"Backend"};
+    LOG(log.info()) << "Constructing BackendInterface";
 
     auto const readOnly = config.valueOr("read_only", false);
 
@@ -42,8 +49,7 @@ make_Backend(boost::asio::io_context& ioc, clio::Config const& config)
     if (boost::iequals(type, "cassandra") or boost::iequals(type, "cassandra-new"))
     {
         auto cfg = config.section("database." + type);
-        backend =
-            std::make_shared<Backend::Cassandra::CassandraBackend>(Backend::Cassandra::SettingsProvider{cfg}, readOnly);
+        backend = std::make_shared<data::cassandra::CassandraBackend>(data::cassandra::SettingsProvider{cfg}, readOnly);
     }
 
     if (!backend)
@@ -56,7 +62,7 @@ make_Backend(boost::asio::io_context& ioc, clio::Config const& config)
         backend->updateRange(rng->maxSequence);
     }
 
-    log.info() << "Constructed BackendInterface Successfully";
+    LOG(log.info()) << "Constructed BackendInterface Successfully";
     return backend;
 }
-}  // namespace Backend
+}  // namespace data

src/data/BackendInterface.cpp

@@ -17,28 +17,26 @@
 */
 //==============================================================================
 
-#include <backend/BackendInterface.h>
-#include <log/Logger.h>
+#include <data/BackendInterface.h>
+#include <util/log/Logger.h>
 
 #include <ripple/protocol/Indexes.h>
 #include <ripple/protocol/STLedgerEntry.h>
 
-using namespace clio;
-
 // local to compilation unit loggers
 namespace {
-clio::Logger gLog{"Backend"};
+util::Logger gLog{"Backend"};
 }  // namespace
 
-namespace Backend {
+namespace data {
 bool
 BackendInterface::finishWrites(std::uint32_t const ledgerSequence)
 {
-    gLog.debug() << "Want finish writes for " << ledgerSequence;
+    LOG(gLog.debug()) << "Want finish writes for " << ledgerSequence;
     auto commitRes = doFinishWrites();
     if (commitRes)
     {
-        gLog.debug() << "Successfully commited. Updating range now to " << ledgerSequence;
+        LOG(gLog.debug()) << "Successfully commited. Updating range now to " << ledgerSequence;
         updateRange(ledgerSequence);
     }
     return commitRes;
@@ -50,27 +48,9 @@ BackendInterface::writeLedgerObject(std::string&& key, std::uint32_t const seq,
     doWriteLedgerObject(std::move(key), seq, std::move(blob));
 }
 
-std::optional<LedgerRange>
-BackendInterface::hardFetchLedgerRangeNoThrow(boost::asio::yield_context& yield) const
-{
-    gLog.trace() << "called";
-    while (true)
-    {
-        try
-        {
-            return hardFetchLedgerRange(yield);
-        }
-        catch (DatabaseTimeout& t)
-        {
-            ;
-        }
-    }
-}
-
 std::optional<LedgerRange>
 BackendInterface::hardFetchLedgerRangeNoThrow() const
 {
-    gLog.trace() << "called";
     return retryOnTimeout([&]() { return hardFetchLedgerRange(); });
 }
 
@@ -79,22 +59,22 @@ std::optional<Blob>
 BackendInterface::fetchLedgerObject(
     ripple::uint256 const& key,
     std::uint32_t const sequence,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     auto obj = cache_.get(key, sequence);
     if (obj)
     {
-        gLog.trace() << "Cache hit - " << ripple::strHex(key);
+        LOG(gLog.trace()) << "Cache hit - " << ripple::strHex(key);
         return *obj;
     }
     else
     {
-        gLog.trace() << "Cache miss - " << ripple::strHex(key);
+        LOG(gLog.trace()) << "Cache miss - " << ripple::strHex(key);
         auto dbObj = doFetchLedgerObject(key, sequence, yield);
         if (!dbObj)
-            gLog.trace() << "Missed cache and missed in db";
+            LOG(gLog.trace()) << "Missed cache and missed in db";
         else
-            gLog.trace() << "Missed cache but found in db";
+            LOG(gLog.trace()) << "Missed cache but found in db";
         return dbObj;
     }
 }
@@ -103,7 +83,7 @@ std::vector<Blob>
 BackendInterface::fetchLedgerObjects(
     std::vector<ripple::uint256> const& keys,
     std::uint32_t const sequence,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     std::vector<Blob> results;
     results.resize(keys.size());
@@ -116,7 +96,7 @@ BackendInterface::fetchLedgerObjects(
         else
             misses.push_back(keys[i]);
     }
-    gLog.trace() << "Cache hits = " << keys.size() - misses.size() << " - cache misses = " << misses.size();
+    LOG(gLog.trace()) << "Cache hits = " << keys.size() - misses.size() << " - cache misses = " << misses.size();
 
     if (misses.size())
     {
@@ -138,13 +118,13 @@ std::optional<ripple::uint256>
 BackendInterface::fetchSuccessorKey(
     ripple::uint256 key,
     std::uint32_t const ledgerSequence,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     auto succ = cache_.getSuccessor(key, ledgerSequence);
     if (succ)
-        gLog.trace() << "Cache hit - " << ripple::strHex(key);
+        LOG(gLog.trace()) << "Cache hit - " << ripple::strHex(key);
     else
-        gLog.trace() << "Cache miss - " << ripple::strHex(key);
+        LOG(gLog.trace()) << "Cache miss - " << ripple::strHex(key);
     return succ ? succ->key : doFetchSuccessorKey(key, ledgerSequence, yield);
 }
 
@@ -152,7 +132,7 @@ std::optional<LedgerObject>
 BackendInterface::fetchSuccessorObject(
     ripple::uint256 key,
     std::uint32_t const ledgerSequence,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     auto succ = fetchSuccessorKey(key, ledgerSequence, yield);
     if (succ)
@@ -171,7 +151,7 @@ BackendInterface::fetchBookOffers(
     ripple::uint256 const& book,
     std::uint32_t const ledgerSequence,
     std::uint32_t const limit,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     // TODO try to speed this up. This can take a few seconds. The goal is
     // to get it down to a few hundred milliseconds.
@@ -194,7 +174,7 @@ BackendInterface::fetchBookOffers(
         succMillis += getMillis(mid2 - mid1);
         if (!offerDir || offerDir->key >= bookEnd)
         {
-            gLog.trace() << "offerDir.has_value() " << offerDir.has_value() << " breaking";
+            LOG(gLog.trace()) << "offerDir.has_value() " << offerDir.has_value() << " breaking";
             break;
         }
         uTipIndex = offerDir->key;
@@ -207,7 +187,7 @@ BackendInterface::fetchBookOffers(
             auto next = sle.getFieldU64(ripple::sfIndexNext);
             if (!next)
             {
-                gLog.trace() << "Next is empty. breaking";
+                LOG(gLog.trace()) << "Next is empty. breaking";
                 break;
             }
             auto nextKey = ripple::keylet::page(uTipIndex, next);
@@ -223,32 +203,56 @@ BackendInterface::fetchBookOffers(
     auto objs = fetchLedgerObjects(keys, ledgerSequence, yield);
     for (size_t i = 0; i < keys.size() && i < limit; ++i)
     {
-        gLog.trace() << "Key = " << ripple::strHex(keys[i]) << " blob = " << ripple::strHex(objs[i])
-                     << " ledgerSequence = " << ledgerSequence;
+        LOG(gLog.trace()) << "Key = " << ripple::strHex(keys[i]) << " blob = " << ripple::strHex(objs[i])
+                          << " ledgerSequence = " << ledgerSequence;
         assert(objs[i].size());
         page.offers.push_back({keys[i], objs[i]});
     }
     auto end = std::chrono::system_clock::now();
-    gLog.debug() << "Fetching " << std::to_string(keys.size()) << " offers took "
-                 << std::to_string(getMillis(mid - begin)) << " milliseconds. Fetching next dir took "
-                 << std::to_string(succMillis) << " milliseonds. Fetched next dir " << std::to_string(numSucc)
-                 << " times"
-                 << " Fetching next page of dir took " << std::to_string(pageMillis) << " milliseconds"
-                 << ". num pages = " << std::to_string(numPages) << ". Fetching all objects took "
-                 << std::to_string(getMillis(end - mid))
-                 << " milliseconds. total time = " << std::to_string(getMillis(end - begin)) << " milliseconds"
-                 << " book = " << ripple::strHex(book);
+    LOG(gLog.debug()) << "Fetching " << std::to_string(keys.size()) << " offers took "
+                      << std::to_string(getMillis(mid - begin)) << " milliseconds. Fetching next dir took "
+                      << std::to_string(succMillis) << " milliseonds. Fetched next dir " << std::to_string(numSucc)
+                      << " times"
+                      << " Fetching next page of dir took " << std::to_string(pageMillis) << " milliseconds"
+                      << ". num pages = " << std::to_string(numPages) << ". Fetching all objects took "
+                      << std::to_string(getMillis(end - mid))
+                      << " milliseconds. total time = " << std::to_string(getMillis(end - begin)) << " milliseconds"
+                      << " book = " << ripple::strHex(book);
 
     return page;
 }
 
+std::optional<LedgerRange>
+BackendInterface::hardFetchLedgerRange() const
+{
+    return synchronous([this](auto yield) { return hardFetchLedgerRange(yield); });
+}
+
+std::optional<LedgerRange>
+BackendInterface::fetchLedgerRange() const
+{
+    std::shared_lock lck(rngMtx_);
+    return range;
+}
+
+void
+BackendInterface::updateRange(uint32_t newMax)
+{
+    std::scoped_lock lck(rngMtx_);
+    assert(!range || newMax >= range->maxSequence);
+    if (!range)
+        range = {newMax, newMax};
+    else
+        range->maxSequence = newMax;
+}
+
 LedgerPage
 BackendInterface::fetchLedgerPage(
     std::optional<ripple::uint256> const& cursor,
     std::uint32_t const ledgerSequence,
     std::uint32_t const limit,
     bool outOfOrder,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     LedgerPage page;
 
@@ -272,14 +276,14 @@ BackendInterface::fetchLedgerPage(
             page.objects.push_back({std::move(keys[i]), std::move(objects[i])});
         else if (!outOfOrder)
         {
-            gLog.error() << "Deleted or non-existent object in successor table. key = " << ripple::strHex(keys[i])
-                         << " - seq = " << ledgerSequence;
+            LOG(gLog.error()) << "Deleted or non-existent object in successor table. key = " << ripple::strHex(keys[i])
+                              << " - seq = " << ledgerSequence;
             std::stringstream msg;
             for (size_t j = 0; j < objects.size(); ++j)
             {
                 msg << " - " << ripple::strHex(keys[j]);
             }
-            gLog.error() << msg.str();
+            LOG(gLog.error()) << msg.str();
         }
     }
     if (keys.size() && !reachedEnd)
@@ -289,7 +293,7 @@ BackendInterface::fetchLedgerPage(
 }
 
 std::optional<ripple::Fees>
-BackendInterface::fetchFees(std::uint32_t const seq, boost::asio::yield_context& yield) const
+BackendInterface::fetchFees(std::uint32_t const seq, boost::asio::yield_context yield) const
 {
     ripple::Fees fees;
 
@@ -298,7 +302,7 @@ BackendInterface::fetchFees(std::uint32_t const seq, boost::asio::yield_context&
 
     if (!bytes)
     {
-        gLog.error() << "Could not find fees";
+        LOG(gLog.error()) << "Could not find fees";
         return {};
     }
 
@@ -308,9 +312,6 @@ BackendInterface::fetchFees(std::uint32_t const seq, boost::asio::yield_context&
     if (sle.getFieldIndex(ripple::sfBaseFee) != -1)
         fees.base = sle.getFieldU64(ripple::sfBaseFee);
 
-    if (sle.getFieldIndex(ripple::sfReferenceFeeUnits) != -1)
-        fees.units = sle.getFieldU32(ripple::sfReferenceFeeUnits);
-
     if (sle.getFieldIndex(ripple::sfReserveBase) != -1)
         fees.reserve = sle.getFieldU32(ripple::sfReserveBase);
 
@@ -320,4 +321,4 @@ BackendInterface::fetchFees(std::uint32_t const seq, boost::asio::yield_context&
     return fees;
 }
 
-}  // namespace Backend
+}  // namespace data

src/data/BackendInterface.h

@@ -0,0 +1,568 @@
+//------------------------------------------------------------------------------
+/*
+    This file is part of clio: https://github.com/XRPLF/clio
+    Copyright (c) 2022, the clio developers.
+
+    Permission to use, copy, modify, and distribute this software for any
+    purpose with or without fee is hereby granted, provided that the above
+    copyright notice and this permission notice appear in all copies.
+
+    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
+    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
+    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+//==============================================================================
+
+#pragma once
+
+#include <data/DBHelpers.h>
+#include <data/LedgerCache.h>
+#include <data/Types.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
+
+#include <ripple/protocol/Fees.h>
+#include <ripple/protocol/LedgerHeader.h>
+#include <boost/asio/spawn.hpp>
+#include <boost/json.hpp>
+
+#include <thread>
+#include <type_traits>
+
+namespace data {
+
+/**
+ * @brief Represents a database timeout error.
+ */
+class DatabaseTimeout : public std::exception
+{
+public:
+    const char*
+    what() const throw() override
+    {
+        return "Database read timed out. Please retry the request";
+    }
+};
+
+/**
+ * @brief A helper function that catches DatabaseTimout exceptions and retries indefinitely.
+ *
+ * @tparam FnType The type of function object to execute
+ * @param func The function object to execute
+ * @param waitMs Delay between retry attempts
+ * @return auto The same as the return type of func
+ */
+template <class FnType>
+auto
+retryOnTimeout(FnType func, size_t waitMs = 500)
+{
+    static util::Logger log{"Backend"};
+
+    while (true)
+    {
+        try
+        {
+            return func();
+        }
+        catch (DatabaseTimeout const&)
+        {
+            LOG(log.error()) << "Database request timed out. Sleeping and retrying ... ";
+            std::this_thread::sleep_for(std::chrono::milliseconds(waitMs));
+        }
+    }
+}
+
+/**
+ * @brief Synchronously executes the given function object inside a coroutine.
+ *
+ * @tparam FnType The type of function object to execute
+ * @param func The function object to execute
+ * @return auto The same as the return type of func
+ */
+template <class FnType>
+auto
+synchronous(FnType&& func)
+{
+    boost::asio::io_context ctx;
+
+    using R = typename boost::result_of<FnType(boost::asio::yield_context)>::type;
+    if constexpr (!std::is_same<R, void>::value)
+    {
+        R res;
+        boost::asio::spawn(
+            ctx, [_ = boost::asio::make_work_guard(ctx), &func, &res](auto yield) { res = func(yield); });
+
+        ctx.run();
+        return res;
+    }
+    else
+    {
+        boost::asio::spawn(ctx, [_ = boost::asio::make_work_guard(ctx), &func](auto yield) { func(yield); });
+        ctx.run();
+    }
+}
+
+/**
+ * @brief Synchronously execute the given function object and retry until no DatabaseTimeout is thrown.
+ *
+ * @tparam FnType The type of function object to execute
+ * @param func The function object to execute
+ * @return auto The same as the return type of func
+ */
+template <class FnType>
+auto
+synchronousAndRetryOnTimeout(FnType&& func)
+{
+    return retryOnTimeout([&]() { return synchronous(func); });
+}
+
+/**
+ * @brief The interface to the database used by Clio.
+ */
+class BackendInterface
+{
+protected:
+    mutable std::shared_mutex rngMtx_;
+    std::optional<LedgerRange> range;
+    LedgerCache cache_;
+
+public:
+    BackendInterface() = default;
+    virtual ~BackendInterface() = default;
+
+    // TODO: Remove this hack. Cache should not be exposed thru BackendInterface
+    /**
+     * @return Immutable cache
+     */
+    LedgerCache const&
+    cache() const
+    {
+        return cache_;
+    }
+
+    /**
+     * @return Mutable cache
+     */
+    LedgerCache&
+    cache()
+    {
+        return cache_;
+    }
+
+    /**
+     * @brief Fetches a specific ledger by sequence number.
+     *
+     * @param sequence The sequence number to fetch for
+     * @param yield The coroutine context
+     * @return The ripple::LedgerHeader if found; nullopt otherwise
+     */
+    virtual std::optional<ripple::LedgerHeader>
+    fetchLedgerBySequence(std::uint32_t const sequence, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches a specific ledger by hash.
+     *
+     * @param hash The hash to fetch for
+     * @param yield The coroutine context
+     * @return The ripple::LedgerHeader if found; nullopt otherwise
+     */
+    virtual std::optional<ripple::LedgerHeader>
+    fetchLedgerByHash(ripple::uint256 const& hash, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches the latest ledger sequence.
+     *
+     * @param yield The coroutine context
+     * @return Latest sequence wrapped in an optional if found; nullopt otherwise
+     */
+    virtual std::optional<std::uint32_t>
+    fetchLatestLedgerSequence(boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetch the current ledger range.
+     *
+     * @return The current ledger range if populated; nullopt otherwise
+     */
+    std::optional<LedgerRange>
+    fetchLedgerRange() const;
+
+    /**
+     * @brief Updates the range of sequences that are stored in the DB.
+     *
+     * @param newMax The new maximum sequence available
+     */
+    void
+    updateRange(uint32_t newMax);
+
+    /**
+     * @brief Fetch the fees from a specific ledger sequence.
+     *
+     * @param seq The sequence to fetch for
+     * @param yield The coroutine context
+     * @return ripple::Fees if fees are found; nullopt otherwise
+     */
+    std::optional<ripple::Fees>
+    fetchFees(std::uint32_t const seq, boost::asio::yield_context yield) const;
+
+    /**
+     * @brief Fetches a specific transaction.
+     *
+     * @param hash The hash of the transaction to fetch
+     * @param yield The coroutine context
+     * @return TransactionAndMetadata if transaction is found; nullopt otherwise
+     */
+    virtual std::optional<TransactionAndMetadata>
+    fetchTransaction(ripple::uint256 const& hash, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches multiple transactions.
+     *
+     * @param hashes A vector of hashes to fetch transactions for
+     * @param yield The coroutine context
+     * @return A vector of TransactionAndMetadata matching the given hashes
+     */
+    virtual std::vector<TransactionAndMetadata>
+    fetchTransactions(std::vector<ripple::uint256> const& hashes, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches all transactions for a specific account.
+     *
+     * @param account The account to fetch transactions for
+     * @param limit The maximum number of transactions per result page
+     * @param forward Whether to fetch the page forwards or backwards from the given cursor
+     * @param cursor The cursor to resume fetching from
+     * @param yield The coroutine context
+     * @return Results and a cursor to resume from
+     */
+    virtual TransactionsAndCursor
+    fetchAccountTransactions(
+        ripple::AccountID const& account,
+        std::uint32_t const limit,
+        bool forward,
+        std::optional<TransactionsCursor> const& cursor,
+        boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches all transactions from a specific ledger.
+     *
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return Results as a vector of TransactionAndMetadata
+     */
+    virtual std::vector<TransactionAndMetadata>
+    fetchAllTransactionsInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches all transaction hashes from a specific ledger.
+     *
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return Hashes as ripple::uint256 in a vector
+     */
+    virtual std::vector<ripple::uint256>
+    fetchAllTransactionHashesInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches a specific NFT.
+     *
+     * @param tokenID The ID of the NFT
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return NFT object on success; nullopt otherwise
+     */
+    virtual std::optional<NFT>
+    fetchNFT(ripple::uint256 const& tokenID, std::uint32_t const ledgerSequence, boost::asio::yield_context yield)
+        const = 0;
+
+    /**
+     * @brief Fetches all transactions for a specific NFT.
+     *
+     * @param tokenID The ID of the NFT
+     * @param limit The maximum number of transactions per result page
+     * @param forward Whether to fetch the page forwards or backwards from the given cursor
+     * @param cursorIn The cursor to resume fetching from
+     * @param yield The coroutine context
+     * @return Results and a cursor to resume from
+     */
+    virtual TransactionsAndCursor
+    fetchNFTTransactions(
+        ripple::uint256 const& tokenID,
+        std::uint32_t const limit,
+        bool const forward,
+        std::optional<TransactionsCursor> const& cursorIn,
+        boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches a specific ledger object.
+     *
+     * Currently the real fetch happens in doFetchLedgerObject and fetchLedgerObject attempts to fetch from Cache first
+     * and only calls out to the real DB if a cache miss ocurred.
+     *
+     * @param key The key of the object
+     * @param sequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return The object as a Blob on success; nullopt otherwise
+     */
+    std::optional<Blob>
+    fetchLedgerObject(ripple::uint256 const& key, std::uint32_t const sequence, boost::asio::yield_context yield) const;
+
+    /**
+     * @brief Fetches all ledger objects by their keys.
+     *
+     * Currently the real fetch happens in doFetchLedgerObjects and fetchLedgerObjects attempts to fetch from Cache
+     * first and only calls out to the real DB for each of the keys that was not found in the cache.
+     *
+     * @param keys A vector with the keys of the objects to fetch
+     * @param sequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return A vector of ledger objects as Blobs
+     */
+    std::vector<Blob>
+    fetchLedgerObjects(
+        std::vector<ripple::uint256> const& keys,
+        std::uint32_t const sequence,
+        boost::asio::yield_context yield) const;
+
+    /**
+     * @brief The database-specific implementation for fetching a ledger object.
+     *
+     * @param key The key to fetch for
+     * @param sequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return The object as a Blob on success; nullopt otherwise
+     */
+    virtual std::optional<Blob>
+    doFetchLedgerObject(ripple::uint256 const& key, std::uint32_t const sequence, boost::asio::yield_context yield)
+        const = 0;
+
+    /**
+     * @brief The database-specific implementation for fetching ledger objects.
+     *
+     * @param keys The keys to fetch for
+     * @param sequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return A vector of Blobs representing each fetched object
+     */
+    virtual std::vector<Blob>
+    doFetchLedgerObjects(
+        std::vector<ripple::uint256> const& keys,
+        std::uint32_t const sequence,
+        boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Returns the difference between ledgers.
+     *
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return A vector of LedgerObject representing the diff
+     */
+    virtual std::vector<LedgerObject>
+    fetchLedgerDiff(std::uint32_t const ledgerSequence, boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches a page of ledger objects, ordered by key/index.
+     *
+     * @param cursor The cursor to resume fetching from
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param limit The maximum number of transactions per result page
+     * @param outOfOrder If set to true max available sequence is used instead of ledgerSequence
+     * @param yield The coroutine context
+     * @return The ledger page
+     */
+    LedgerPage
+    fetchLedgerPage(
+        std::optional<ripple::uint256> const& cursor,
+        std::uint32_t const ledgerSequence,
+        std::uint32_t const limit,
+        bool outOfOrder,
+        boost::asio::yield_context yield) const;
+
+    /**
+     * @brief Fetches the successor object.
+     *
+     * @param key The key to fetch for
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return The sucessor on success; nullopt otherwise
+     */
+    std::optional<LedgerObject>
+    fetchSuccessorObject(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context yield)
+        const;
+
+    /**
+     * @brief Fetches the successor key.
+     *
+     * Thea real fetch happens in doFetchSuccessorKey. This function will attempt to lookup the successor in the cache
+     * first and only if it's not found in the cache will it fetch from the actual DB.
+     *
+     * @param key The key to fetch for
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return The sucessor key on success; nullopt otherwise
+     */
+    std::optional<ripple::uint256>
+    fetchSuccessorKey(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context yield) const;
+
+    /**
+     * @brief Database-specific implementation of fetching the successor key
+     *
+     * @param key The key to fetch for
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param yield The coroutine context
+     * @return The sucessor on success; nullopt otherwise
+     */
+    virtual std::optional<ripple::uint256>
+    doFetchSuccessorKey(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context yield)
+        const = 0;
+
+    /**
+     * @brief Fetches book offers.
+     *
+     * @param book Unsigned 256-bit integer.
+     * @param ledgerSequence The ledger sequence to fetch for
+     * @param limit Pagaing limit as to how many transactions returned per page.
+     * @param yield The coroutine context
+     * @return The book offers page
+     */
+    BookOffersPage
+    fetchBookOffers(
+        ripple::uint256 const& book,
+        std::uint32_t const ledgerSequence,
+        std::uint32_t const limit,
+        boost::asio::yield_context yield) const;
+
+    /**
+     * @brief Synchronously fetches the ledger range from DB.
+     *
+     * This function just wraps hardFetchLedgerRange(boost::asio::yield_context) using synchronous(FnType&&).
+     *
+     * @return The ledger range if available; nullopt otherwise
+     */
+    std::optional<LedgerRange>
+    hardFetchLedgerRange() const;
+
+    /**
+     * @brief Fetches the ledger range from DB.
+     *
+     * @return The ledger range if available; nullopt otherwise
+     */
+    virtual std::optional<LedgerRange>
+    hardFetchLedgerRange(boost::asio::yield_context yield) const = 0;
+
+    /**
+     * @brief Fetches the ledger range from DB retrying until no DatabaseTimeout is thrown.
+     *
+     * @return The ledger range if available; nullopt otherwise
+     */
+    std::optional<LedgerRange>
+    hardFetchLedgerRangeNoThrow() const;
+
+    /**
+     * @brief Writes to a specific ledger.
+     *
+     * @param ledgerHeader Ledger header.
+     * @param blob r-value string serialization of ledger header.
+     */
+    virtual void
+    writeLedger(ripple::LedgerHeader const& ledgerHeader, std::string&& blob) = 0;
+
+    /**
+     * @brief Writes a new ledger object.
+     *
+     * @param key The key to write the ledger object under
+     * @param seq The ledger sequence to write for
+     * @param blob The data to write
+     */
+    virtual void
+    writeLedgerObject(std::string&& key, std::uint32_t const seq, std::string&& blob);
+
+    /**
+     * @brief Writes a new transaction.
+     *
+     * @param hash The hash of the transaction
+     * @param seq The ledger sequence to write for
+     * @param date The timestamp of the entry
+     * @param transaction The transaction data to write
+     * @param metadata The metadata to write
+     */
+    virtual void
+    writeTransaction(
+        std::string&& hash,
+        std::uint32_t const seq,
+        std::uint32_t const date,
+        std::string&& transaction,
+        std::string&& metadata) = 0;
+
+    /**
+     * @brief Writes NFTs to the database.
+     *
+     * @param data A vector of NFTsData objects representing the NFTs
+     */
+    virtual void
+    writeNFTs(std::vector<NFTsData>&& data) = 0;
+
+    /**
+     * @brief Write a new set of account transactions.
+     *
+     * @param data A vector of AccountTransactionsData objects representing the account transactions
+     */
+    virtual void
+    writeAccountTransactions(std::vector<AccountTransactionsData>&& data) = 0;
+
+    /**
+     * @brief Write NFTs transactions.
+     *
+     * @param data A vector of NFTTransactionsData objects
+     */
+    virtual void
+    writeNFTTransactions(std::vector<NFTTransactionsData>&& data) = 0;
+
+    /**
+     * @brief Write a new successor.
+     *
+     * @param key Key of the object that the passed successor will be the successor for
+     * @param seq The ledger sequence to write for
+     * @param successor The successor data to write
+     */
+    virtual void
+    writeSuccessor(std::string&& key, std::uint32_t const seq, std::string&& successor) = 0;
+
+    /**
+     * @brief Starts a write transaction with the DB. No-op for cassandra.
+     *
+     * Note: Can potentially be deprecated and removed.
+     */
+    virtual void
+    startWrites() const = 0;
+
+    /**
+     * @brief Tells database we finished writing all data for a specific ledger.
+     *
+     * Uses doFinishWrites to synchronize with the pending writes.
+     *
+     * @param ledgerSequence The ledger sequence to finish writing for
+     * @return true on success; false otherwise
+     */
+    bool
+    finishWrites(std::uint32_t const ledgerSequence);
+
+    /**
+     * @return true if database is overwhelmed; false otherwise
+     */
+    virtual bool
+    isTooBusy() const = 0;
+
+private:
+    virtual void
+    doWriteLedgerObject(std::string&& key, std::uint32_t const seq, std::string&& blob) = 0;
+
+    virtual bool
+    doFinishWrites() = 0;
+};
+
+}  // namespace data
+using BackendInterface = data::BackendInterface;

src/data/CassandraBackend.h

@@ -19,41 +19,41 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
-#include <backend/cassandra/Concepts.h>
-#include <backend/cassandra/Handle.h>
-#include <backend/cassandra/Schema.h>
-#include <backend/cassandra/SettingsProvider.h>
-#include <backend/cassandra/impl/ExecutionStrategy.h>
-#include <log/Logger.h>
+#include <data/BackendInterface.h>
+#include <data/cassandra/Concepts.h>
+#include <data/cassandra/Handle.h>
+#include <data/cassandra/Schema.h>
+#include <data/cassandra/SettingsProvider.h>
+#include <data/cassandra/impl/ExecutionStrategy.h>
 #include <util/LedgerUtils.h>
 #include <util/Profiler.h>
+#include <util/log/Logger.h>
 
-#include <ripple/app/tx/impl/details/NFTokenUtils.h>
+#include <ripple/protocol/LedgerHeader.h>
+#include <ripple/protocol/nft.h>
 #include <boost/asio/spawn.hpp>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 /**
- * @brief Implements @ref BackendInterface for Cassandra/Scylladb
+ * @brief Implements @ref BackendInterface for Cassandra/ScyllaDB.
  *
- * Note: this is a safer and more correct rewrite of the original implementation
- * of the backend. We deliberately did not change the interface for now so that
- * other parts such as ETL do not have to change at all.
- * Eventually we should change the interface so that it does not have to know
- * about yield_context.
+ * Note: This is a safer and more correct rewrite of the original implementation of the backend.
+ *
+ * @tparam SettingsProviderType The settings provider type to use
+ * @tparam ExecutionStrategyType The execution strategy type to use
  */
-template <SomeSettingsProvider SettingsProviderType, SomeExecutionStrategy ExecutionStrategy>
+template <SomeSettingsProvider SettingsProviderType, SomeExecutionStrategy ExecutionStrategyType>
 class BasicCassandraBackend : public BackendInterface
 {
-    clio::Logger log_{"Backend"};
+    util::Logger log_{"Backend"};
 
     SettingsProviderType settingsProvider_;
     Schema<SettingsProviderType> schema_;
     Handle handle_;
 
     // have to be mutable because BackendInterface constness :(
-    mutable ExecutionStrategy executor_;
+    mutable ExecutionStrategyType executor_;
 
     std::atomic_uint32_t ledgerSequence_ = 0u;
 
@@ -61,7 +61,8 @@ public:
     /**
      * @brief Create a new cassandra/scylla backend instance.
      *
-     * @param settingsProvider
+     * @param settingsProvider The settings provider to use
+     * @param readOnly Whether the database should be in readonly mode
      */
     BasicCassandraBackend(SettingsProviderType settingsProvider, bool readOnly)
         : settingsProvider_{std::move(settingsProvider)}
@@ -92,11 +93,11 @@ public:
         }
         catch (std::runtime_error const& ex)
         {
-            log_.error() << "Failed to prepare the statements: " << ex.what() << "; readOnly: " << readOnly;
+            LOG(log_.error()) << "Failed to prepare the statements: " << ex.what() << "; readOnly: " << readOnly;
             throw;
         }
 
-        log_.info() << "Created (revamped) CassandraBackend";
+        LOG(log_.info()) << "Created (revamped) CassandraBackend";
     }
 
     TransactionsAndCursor
@@ -105,7 +106,7 @@ public:
         std::uint32_t const limit,
         bool forward,
         std::optional<TransactionsCursor> const& cursorIn,
-        boost::asio::yield_context& yield) const override
+        boost::asio::yield_context yield) const override
     {
         auto rng = fetchLedgerRange();
         if (!rng)
@@ -122,8 +123,8 @@ public:
         if (cursor)
         {
             statement.bindAt(1, cursor->asTuple());
-            log_.debug() << "account = " << ripple::strHex(account) << " tuple = " << cursor->ledgerSequence
-                         << cursor->transactionIndex;
+            LOG(log_.debug()) << "account = " << ripple::strHex(account) << " tuple = " << cursor->ledgerSequence
+                              << cursor->transactionIndex;
         }
         else
         {
@@ -131,7 +132,8 @@ public:
             auto const placeHolder = forward ? 0u : std::numeric_limits<std::uint32_t>::max();
 
             statement.bindAt(1, std::make_tuple(placeHolder, placeHolder));
-            log_.debug() << "account = " << ripple::strHex(account) << " idx = " << seq << " tuple = " << placeHolder;
+            LOG(log_.debug()) << "account = " << ripple::strHex(account) << " idx = " << seq
+                              << " tuple = " << placeHolder;
         }
 
         // FIXME: Limit is a hack to support uint32_t properly for the time
@@ -142,20 +144,20 @@ public:
         auto const& results = res.value();
         if (not results.hasRows())
         {
-            log_.debug() << "No rows returned";
+            LOG(log_.debug()) << "No rows returned";
             return {};
         }
 
         std::vector<ripple::uint256> hashes = {};
         auto numRows = results.numRows();
-        log_.info() << "num_rows = " << numRows;
+        LOG(log_.info()) << "num_rows = " << numRows;
 
         for (auto [hash, data] : extract<ripple::uint256, std::tuple<uint32_t, uint32_t>>(results))
         {
             hashes.push_back(hash);
             if (--numRows == 0)
             {
-                log_.debug() << "Setting cursor";
+                LOG(log_.debug()) << "Setting cursor";
                 cursor = data;
 
                 // forward queries by ledger/tx sequence `>=`
@@ -166,11 +168,11 @@ public:
         }
 
         auto const txns = fetchTransactions(hashes, yield);
-        log_.debug() << "Txns = " << txns.size();
+        LOG(log_.debug()) << "Txns = " << txns.size();
 
         if (txns.size() == limit)
         {
-            log_.debug() << "Returning cursor";
+            LOG(log_.debug()) << "Returning cursor";
             return {txns, cursor};
         }
 
@@ -190,18 +192,18 @@ public:
 
         if (not executeSyncUpdate(schema_->updateLedgerRange.bind(ledgerSequence_, true, ledgerSequence_ - 1)))
         {
-            log_.warn() << "Update failed for ledger " << ledgerSequence_;
+            LOG(log_.warn()) << "Update failed for ledger " << ledgerSequence_;
             return false;
         }
 
-        log_.info() << "Committed ledger " << ledgerSequence_;
+        LOG(log_.info()) << "Committed ledger " << ledgerSequence_;
         return true;
     }
 
     void
-    writeLedger(ripple::LedgerInfo const& ledgerInfo, std::string&& header) override
+    writeLedger(ripple::LedgerHeader const& ledgerInfo, std::string&& blob) override
     {
-        executor_.write(schema_->insertLedgerHeader, ledgerInfo.seq, std::move(header));
+        executor_.write(schema_->insertLedgerHeader, ledgerInfo.seq, std::move(blob));
 
         executor_.write(schema_->insertLedgerHash, ledgerInfo.hash, ledgerInfo.seq);
 
@@ -209,7 +211,7 @@ public:
     }
 
     std::optional<std::uint32_t>
-    fetchLatestLedgerSequence(boost::asio::yield_context& yield) const override
+    fetchLatestLedgerSequence(boost::asio::yield_context yield) const override
     {
         if (auto const res = executor_.read(yield, schema_->selectLatestLedger); res)
         {
@@ -218,25 +220,23 @@ public:
                 if (auto const maybeValue = result.template get<uint32_t>(); maybeValue)
                     return maybeValue;
 
-                log_.error() << "Could not fetch latest ledger - no rows";
+                LOG(log_.error()) << "Could not fetch latest ledger - no rows";
                 return std::nullopt;
             }
 
-            log_.error() << "Could not fetch latest ledger - no result";
+            LOG(log_.error()) << "Could not fetch latest ledger - no result";
         }
         else
         {
-            log_.error() << "Could not fetch latest ledger: " << res.error();
+            LOG(log_.error()) << "Could not fetch latest ledger: " << res.error();
         }
 
         return std::nullopt;
     }
 
-    std::optional<ripple::LedgerInfo>
-    fetchLedgerBySequence(std::uint32_t const sequence, boost::asio::yield_context& yield) const override
+    std::optional<ripple::LedgerHeader>
+    fetchLedgerBySequence(std::uint32_t const sequence, boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call for seq " << sequence;
-
         auto const res = executor_.read(yield, schema_->selectLedgerBySeq, sequence);
         if (res)
         {
@@ -247,25 +247,23 @@ public:
                     return util::deserializeHeader(ripple::makeSlice(*maybeValue));
                 }
 
-                log_.error() << "Could not fetch ledger by sequence - no rows";
+                LOG(log_.error()) << "Could not fetch ledger by sequence - no rows";
                 return std::nullopt;
             }
 
-            log_.error() << "Could not fetch ledger by sequence - no result";
+            LOG(log_.error()) << "Could not fetch ledger by sequence - no result";
         }
         else
         {
-            log_.error() << "Could not fetch ledger by sequence: " << res.error();
+            LOG(log_.error()) << "Could not fetch ledger by sequence: " << res.error();
         }
 
         return std::nullopt;
     }
 
-    std::optional<ripple::LedgerInfo>
-    fetchLedgerByHash(ripple::uint256 const& hash, boost::asio::yield_context& yield) const override
+    std::optional<ripple::LedgerHeader>
+    fetchLedgerByHash(ripple::uint256 const& hash, boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
         if (auto const res = executor_.read(yield, schema_->selectLedgerByHash, hash); res)
         {
             if (auto const& result = res.value(); result)
@@ -273,31 +271,29 @@ public:
                 if (auto const maybeValue = result.template get<uint32_t>(); maybeValue)
                     return fetchLedgerBySequence(*maybeValue, yield);
 
-                log_.error() << "Could not fetch ledger by hash - no rows";
+                LOG(log_.error()) << "Could not fetch ledger by hash - no rows";
                 return std::nullopt;
             }
 
-            log_.error() << "Could not fetch ledger by hash - no result";
+            LOG(log_.error()) << "Could not fetch ledger by hash - no result";
         }
         else
         {
-            log_.error() << "Could not fetch ledger by hash: " << res.error();
+            LOG(log_.error()) << "Could not fetch ledger by hash: " << res.error();
         }
 
         return std::nullopt;
     }
 
     std::optional<LedgerRange>
-    hardFetchLedgerRange(boost::asio::yield_context& yield) const override
+    hardFetchLedgerRange(boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
         if (auto const res = executor_.read(yield, schema_->selectLedgerRange); res)
         {
             auto const& results = res.value();
             if (not results.hasRows())
             {
-                log_.debug() << "Could not fetch ledger range - no rows";
+                LOG(log_.debug()) << "Could not fetch ledger range - no rows";
                 return std::nullopt;
             }
 
@@ -319,44 +315,43 @@ public:
             if (range.minSequence > range.maxSequence)
                 std::swap(range.minSequence, range.maxSequence);
 
-            log_.debug() << "After hardFetchLedgerRange range is " << range.minSequence << ":" << range.maxSequence;
+            LOG(log_.debug()) << "After hardFetchLedgerRange range is " << range.minSequence << ":"
+                              << range.maxSequence;
             return range;
         }
         else
         {
-            log_.error() << "Could not fetch ledger range: " << res.error();
+            LOG(log_.error()) << "Could not fetch ledger range: " << res.error();
         }
 
         return std::nullopt;
     }
 
     std::vector<TransactionAndMetadata>
-    fetchAllTransactionsInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context& yield) const override
+    fetchAllTransactionsInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
         auto hashes = fetchAllTransactionHashesInLedger(ledgerSequence, yield);
         return fetchTransactions(hashes, yield);
     }
 
     std::vector<ripple::uint256>
-    fetchAllTransactionHashesInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context& yield)
+    fetchAllTransactionHashesInLedger(std::uint32_t const ledgerSequence, boost::asio::yield_context yield)
         const override
     {
-        log_.trace() << __func__ << " call";
         auto start = std::chrono::system_clock::now();
         auto const res = executor_.read(yield, schema_->selectAllTransactionHashesInLedger, ledgerSequence);
 
         if (not res)
         {
-            log_.error() << "Could not fetch all transaction hashes: " << res.error();
+            LOG(log_.error()) << "Could not fetch all transaction hashes: " << res.error();
             return {};
         }
 
         auto const& result = res.value();
         if (not result.hasRows())
         {
-            log_.error() << "Could not fetch all transaction hashes - no rows; ledger = "
-                         << std::to_string(ledgerSequence);
+            LOG(log_.error()) << "Could not fetch all transaction hashes - no rows; ledger = "
+                              << std::to_string(ledgerSequence);
             return {};
         }
 
@@ -365,18 +360,17 @@ public:
             hashes.push_back(std::move(hash));
 
         auto end = std::chrono::system_clock::now();
-        log_.debug() << "Fetched " << hashes.size() << " transaction hashes from Cassandra in "
-                     << std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count() << " milliseconds";
+        LOG(log_.debug()) << "Fetched " << hashes.size() << " transaction hashes from Cassandra in "
+                          << std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count()
+                          << " milliseconds";
 
         return hashes;
     }
 
     std::optional<NFT>
-    fetchNFT(ripple::uint256 const& tokenID, std::uint32_t const ledgerSequence, boost::asio::yield_context& yield)
+    fetchNFT(ripple::uint256 const& tokenID, std::uint32_t const ledgerSequence, boost::asio::yield_context yield)
         const override
     {
-        log_.trace() << __func__ << " call";
-
         auto const res = executor_.read(yield, schema_->selectNFT, tokenID, ledgerSequence);
         if (not res)
             return std::nullopt;
@@ -407,7 +401,7 @@ public:
             return result;
         }
 
-        log_.error() << "Could not fetch NFT - no rows";
+        LOG(log_.error()) << "Could not fetch NFT - no rows";
         return std::nullopt;
     }
 
@@ -417,10 +411,8 @@ public:
         std::uint32_t const limit,
         bool const forward,
         std::optional<TransactionsCursor> const& cursorIn,
-        boost::asio::yield_context& yield) const override
+        boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
         auto rng = fetchLedgerRange();
         if (!rng)
             return {{}, {}};
@@ -436,8 +428,8 @@ public:
         if (cursor)
         {
             statement.bindAt(1, cursor->asTuple());
-            log_.debug() << "token_id = " << ripple::strHex(tokenID) << " tuple = " << cursor->ledgerSequence
-                         << cursor->transactionIndex;
+            LOG(log_.debug()) << "token_id = " << ripple::strHex(tokenID) << " tuple = " << cursor->ledgerSequence
+                              << cursor->transactionIndex;
         }
         else
         {
@@ -445,7 +437,8 @@ public:
             auto const placeHolder = forward ? 0 : std::numeric_limits<std::uint32_t>::max();
 
             statement.bindAt(1, std::make_tuple(placeHolder, placeHolder));
-            log_.debug() << "token_id = " << ripple::strHex(tokenID) << " idx = " << seq << " tuple = " << placeHolder;
+            LOG(log_.debug()) << "token_id = " << ripple::strHex(tokenID) << " idx = " << seq
+                              << " tuple = " << placeHolder;
         }
 
         statement.bindAt(2, Limit{limit});
@@ -454,20 +447,20 @@ public:
         auto const& results = res.value();
         if (not results.hasRows())
         {
-            log_.debug() << "No rows returned";
+            LOG(log_.debug()) << "No rows returned";
             return {};
         }
 
         std::vector<ripple::uint256> hashes = {};
         auto numRows = results.numRows();
-        log_.info() << "num_rows = " << numRows;
+        LOG(log_.info()) << "num_rows = " << numRows;
 
         for (auto [hash, data] : extract<ripple::uint256, std::tuple<uint32_t, uint32_t>>(results))
         {
             hashes.push_back(hash);
             if (--numRows == 0)
             {
-                log_.debug() << "Setting cursor";
+                LOG(log_.debug()) << "Setting cursor";
                 cursor = data;
 
                 // forward queries by ledger/tx sequence `>=`
@@ -478,11 +471,11 @@ public:
         }
 
         auto const txns = fetchTransactions(hashes, yield);
-        log_.debug() << "NFT Txns = " << txns.size();
+        LOG(log_.debug()) << "NFT Txns = " << txns.size();
 
         if (txns.size() == limit)
         {
-            log_.debug() << "Returning cursor";
+            LOG(log_.debug()) << "Returning cursor";
             return {txns, cursor};
         }
 
@@ -490,10 +483,10 @@ public:
     }
 
     std::optional<Blob>
-    doFetchLedgerObject(ripple::uint256 const& key, std::uint32_t const sequence, boost::asio::yield_context& yield)
+    doFetchLedgerObject(ripple::uint256 const& key, std::uint32_t const sequence, boost::asio::yield_context yield)
         const override
     {
-        log_.debug() << "Fetching ledger object for seq " << sequence << ", key = " << ripple::to_string(key);
+        LOG(log_.debug()) << "Fetching ledger object for seq " << sequence << ", key = " << ripple::to_string(key);
         if (auto const res = executor_.read(yield, schema_->selectObject, key, sequence); res)
         {
             if (auto const result = res->template get<Blob>(); result)
@@ -503,22 +496,20 @@ public:
             }
             else
             {
-                log_.debug() << "Could not fetch ledger object - no rows";
+                LOG(log_.debug()) << "Could not fetch ledger object - no rows";
             }
         }
         else
         {
-            log_.error() << "Could not fetch ledger object: " << res.error();
+            LOG(log_.error()) << "Could not fetch ledger object: " << res.error();
         }
 
         return std::nullopt;
     }
 
     std::optional<TransactionAndMetadata>
-    fetchTransaction(ripple::uint256 const& hash, boost::asio::yield_context& yield) const override
+    fetchTransaction(ripple::uint256 const& hash, boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
         if (auto const res = executor_.read(yield, schema_->selectTransaction, hash); res)
         {
             if (auto const maybeValue = res->template get<Blob, Blob, uint32_t, uint32_t>(); maybeValue)
@@ -528,23 +519,21 @@ public:
             }
             else
             {
-                log_.debug() << "Could not fetch transaction - no rows";
+                LOG(log_.debug()) << "Could not fetch transaction - no rows";
             }
         }
         else
         {
-            log_.error() << "Could not fetch transaction: " << res.error();
+            LOG(log_.error()) << "Could not fetch transaction: " << res.error();
         }
 
         return std::nullopt;
     }
 
     std::optional<ripple::uint256>
-    doFetchSuccessorKey(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context& yield)
+    doFetchSuccessorKey(ripple::uint256 key, std::uint32_t const ledgerSequence, boost::asio::yield_context yield)
         const override
     {
-        log_.trace() << __func__ << " call";
-
         if (auto const res = executor_.read(yield, schema_->selectSuccessor, key, ledgerSequence); res)
         {
             if (auto const result = res->template get<ripple::uint256>(); result)
@@ -555,22 +544,20 @@ public:
             }
             else
             {
-                log_.debug() << "Could not fetch successor - no rows";
+                LOG(log_.debug()) << "Could not fetch successor - no rows";
             }
         }
         else
         {
-            log_.error() << "Could not fetch successor: " << res.error();
+            LOG(log_.error()) << "Could not fetch successor: " << res.error();
         }
 
         return std::nullopt;
     }
 
     std::vector<TransactionAndMetadata>
-    fetchTransactions(std::vector<ripple::uint256> const& hashes, boost::asio::yield_context& yield) const override
+    fetchTransactions(std::vector<ripple::uint256> const& hashes, boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
         if (hashes.size() == 0)
             return {};
 
@@ -581,7 +568,7 @@ public:
         std::vector<Statement> statements;
         statements.reserve(numHashes);
 
-        auto const timeDiff = util::timed([this, &yield, &results, &hashes, &statements]() {
+        auto const timeDiff = util::timed([this, yield, &results, &hashes, &statements]() {
             // TODO: seems like a job for "hash IN (list of hashes)" instead?
             std::transform(
                 std::cbegin(hashes), std::cend(hashes), std::back_inserter(statements), [this](auto const& hash) {
@@ -602,7 +589,8 @@ public:
         });
 
         assert(numHashes == results.size());
-        log_.debug() << "Fetched " << numHashes << " transactions from Cassandra in " << timeDiff << " milliseconds";
+        LOG(log_.debug()) << "Fetched " << numHashes << " transactions from Cassandra in " << timeDiff
+                          << " milliseconds";
         return results;
     }
 
@@ -610,15 +598,13 @@ public:
     doFetchLedgerObjects(
         std::vector<ripple::uint256> const& keys,
         std::uint32_t const sequence,
-        boost::asio::yield_context& yield) const override
+        boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
         if (keys.size() == 0)
             return {};
 
         auto const numKeys = keys.size();
-        log_.trace() << "Fetching " << numKeys << " objects";
+        LOG(log_.trace()) << "Fetching " << numKeys << " objects";
 
         std::vector<Blob> results;
         results.reserve(numKeys);
@@ -641,42 +627,41 @@ public:
                     return {};
             });
 
-        log_.trace() << "Fetched " << numKeys << " objects";
+        LOG(log_.trace()) << "Fetched " << numKeys << " objects";
         return results;
     }
 
     std::vector<LedgerObject>
-    fetchLedgerDiff(std::uint32_t const ledgerSequence, boost::asio::yield_context& yield) const override
+    fetchLedgerDiff(std::uint32_t const ledgerSequence, boost::asio::yield_context yield) const override
     {
-        log_.trace() << __func__ << " call";
-
-        auto const [keys, timeDiff] = util::timed([this, &ledgerSequence, &yield]() -> std::vector<ripple::uint256> {
+        auto const [keys, timeDiff] = util::timed([this, &ledgerSequence, yield]() -> std::vector<ripple::uint256> {
             auto const res = executor_.read(yield, schema_->selectDiff, ledgerSequence);
             if (not res)
             {
-                log_.error() << "Could not fetch ledger diff: " << res.error() << "; ledger = " << ledgerSequence;
+                LOG(log_.error()) << "Could not fetch ledger diff: " << res.error() << "; ledger = " << ledgerSequence;
                 return {};
             }
 
             auto const& results = res.value();
             if (not results)
             {
-                log_.error() << "Could not fetch ledger diff - no rows; ledger = " << ledgerSequence;
+                LOG(log_.error()) << "Could not fetch ledger diff - no rows; ledger = " << ledgerSequence;
                 return {};
             }
 
-            std::vector<ripple::uint256> keys;
+            std::vector<ripple::uint256> resultKeys;
             for (auto [key] : extract<ripple::uint256>(results))
-                keys.push_back(key);
+                resultKeys.push_back(key);
 
-            return keys;
+            return resultKeys;
         });
 
         // one of the above errors must have happened
         if (keys.empty())
             return {};
 
-        log_.debug() << "Fetched " << keys.size() << " diff hashes from Cassandra in " << timeDiff << " milliseconds";
+        LOG(log_.debug()) << "Fetched " << keys.size() << " diff hashes from Cassandra in " << timeDiff
+                          << " milliseconds";
 
         auto const objs = fetchLedgerObjects(keys, ledgerSequence, yield);
         std::vector<LedgerObject> results;
@@ -697,7 +682,7 @@ public:
     void
     doWriteLedgerObject(std::string&& key, std::uint32_t const seq, std::string&& blob) override
     {
-        log_.trace() << " Writing ledger object " << key.size() << ":" << seq << " [" << blob.size() << " bytes]";
+        LOG(log_.trace()) << " Writing ledger object " << key.size() << ":" << seq << " [" << blob.size() << " bytes]";
 
         if (range)
             executor_.write(schema_->insertDiff, seq, key);
@@ -708,8 +693,8 @@ public:
     void
     writeSuccessor(std::string&& key, std::uint32_t const seq, std::string&& successor) override
     {
-        log_.trace() << "Writing successor. key = " << key.size() << " bytes. "
-                     << " seq = " << std::to_string(seq) << " successor = " << successor.size() << " bytes.";
+        LOG(log_.trace()) << "Writing successor. key = " << key.size() << " bytes. "
+                          << " seq = " << std::to_string(seq) << " successor = " << successor.size() << " bytes.";
         assert(key.size() != 0);
         assert(successor.size() != 0);
 
@@ -761,7 +746,7 @@ public:
         std::string&& transaction,
         std::string&& metadata) override
     {
-        log_.trace() << "Writing txn to cassandra";
+        LOG(log_.trace()) << "Writing txn to cassandra";
 
         executor_.write(schema_->insertLedgerTransaction, seq, hash);
         executor_.write(
@@ -819,13 +804,13 @@ private:
         auto maybeSuccess = res->template get<bool>();
         if (not maybeSuccess)
         {
-            log_.error() << "executeSyncUpdate - error getting result - no row";
+            LOG(log_.error()) << "executeSyncUpdate - error getting result - no row";
             return false;
         }
 
         if (not maybeSuccess.value())
         {
-            log_.warn() << "Update failed. Checking if DB state is what we expect";
+            LOG(log_.warn()) << "Update failed. Checking if DB state is what we expect";
 
             // error may indicate that another writer wrote something.
             // in this case let's just compare the current state of things
@@ -841,4 +826,4 @@ private:
 
 using CassandraBackend = BasicCassandraBackend<SettingsProvider, detail::DefaultExecutionStrategy<>>;
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/DBHelpers.h

@@ -17,21 +17,21 @@
 */
 //==============================================================================
 
+/** @file */
 #pragma once
 
 #include <ripple/basics/Log.h>
 #include <ripple/basics/StringUtilities.h>
-#include <ripple/ledger/ReadView.h>
 #include <ripple/protocol/SField.h>
 #include <ripple/protocol/STAccount.h>
 #include <ripple/protocol/TxMeta.h>
 
 #include <boost/container/flat_set.hpp>
 
-#include <backend/Types.h>
+#include <data/Types.h>
 
 /**
- * @brief Struct used to keep track of what to write to account_transactions/account_tx tables
+ * @brief Struct used to keep track of what to write to account_transactions/account_tx tables.
  */
 struct AccountTransactionsData
 {
@@ -40,7 +40,7 @@ struct AccountTransactionsData
     std::uint32_t transactionIndex;
     ripple::uint256 txHash;
 
-    AccountTransactionsData(ripple::TxMeta& meta, ripple::uint256 const& txHash, beast::Journal& j)
+    AccountTransactionsData(ripple::TxMeta& meta, ripple::uint256 const& txHash)
         : accounts(meta.getAffectedAccounts())
         , ledgerSequence(meta.getLgrSeq())
         , transactionIndex(meta.getIndex())
@@ -52,7 +52,7 @@ struct AccountTransactionsData
 };
 
 /**
- * @brief Represents a link from a tx to an NFT that was targeted/modified/created by it
+ * @brief Represents a link from a tx to an NFT that was targeted/modified/created by it.
  *
  * Gets written to nf_token_transactions table and the like.
  */
@@ -139,6 +139,12 @@ struct NFTsData
     }
 };
 
+/**
+ * @brief Check whether the supplied object is an offer.
+ *
+ * @param object The object to check
+ * @return true if the object is an offer; false otherwise
+ */
 template <class T>
 inline bool
 isOffer(T const& object)
@@ -147,19 +153,28 @@ isOffer(T const& object)
     return offer_bytes == 0x006f;
 }
 
+/**
+ * @brief Check whether the supplied hex represents an offer object.
+ *
+ * @param object The object to check
+ * @return true if the object is an offer; false otherwise
+ */
 template <class T>
 inline bool
 isOfferHex(T const& object)
 {
     auto blob = ripple::strUnHex(4, object.begin(), object.begin() + 4);
     if (blob)
-    {
-        short offer_bytes = ((*blob)[1] << 8) | (*blob)[2];
-        return offer_bytes == 0x006f;
-    }
+        return isOffer(*blob);
     return false;
 }
 
+/**
+ * @brief Check whether the supplied object is a dir node.
+ *
+ * @param object The object to check
+ * @return true if the object is a dir node; false otherwise
+ */
 template <class T>
 inline bool
 isDirNode(T const& object)
@@ -168,6 +183,13 @@ isDirNode(T const& object)
     return spaceKey == 0x0064;
 }
 
+/**
+ * @brief Check whether the supplied object is a book dir.
+ *
+ * @param key The key into the object
+ * @param object The object to check
+ * @return true if the object is a book dir; false otherwise
+ */
 template <class T, class R>
 inline bool
 isBookDir(T const& key, R const& object)
@@ -179,6 +201,12 @@ isBookDir(T const& key, R const& object)
     return !sle[~ripple::sfOwner].has_value();
 }
 
+/**
+ * @brief Get the book out of an offer object.
+ *
+ * @param offer The offer to get the book for
+ * @return Book as ripple::uint256
+ */
 template <class T>
 inline ripple::uint256
 getBook(T const& offer)
@@ -186,26 +214,40 @@ getBook(T const& offer)
     ripple::SerialIter it{offer.data(), offer.size()};
     ripple::SLE sle{it, {}};
     ripple::uint256 book = sle.getFieldH256(ripple::sfBookDirectory);
+
     return book;
 }
 
+/**
+ * @brief Get the book base.
+ *
+ * @param key The key to get the book base out of
+ * @return Book base as ripple::uint256
+ */
 template <class T>
 inline ripple::uint256
 getBookBase(T const& key)
 {
     assert(key.size() == ripple::uint256::size());
+
     ripple::uint256 ret;
     for (size_t i = 0; i < 24; ++i)
-    {
         ret.data()[i] = key.data()[i];
-    }
+
     return ret;
 }
 
+/**
+ * @brief Stringify a ripple::uint256.
+ *
+ * @param input The input value
+ * @return The input value as a string
+ */
 inline std::string
-uint256ToString(ripple::uint256 const& uint)
+uint256ToString(ripple::uint256 const& input)
 {
-    return {reinterpret_cast<const char*>(uint.data()), uint.size()};
+    return {reinterpret_cast<const char*>(input.data()), input.size()};
 }
 
+/** @brief The ripple epoch start timestamp. Midnight on 1st January 2000. */
 static constexpr std::uint32_t rippleEpochStart = 946684800;

src/data/LedgerCache.cpp

@@ -17,9 +17,9 @@
 */
 //==============================================================================
 
-#include <backend/LedgerCache.h>
+#include <data/LedgerCache.h>
 
-namespace Backend {
+namespace data {
 
 uint32_t
 LedgerCache::latestLedgerSequence() const
@@ -69,7 +69,7 @@ LedgerCache::getSuccessor(ripple::uint256 const& key, uint32_t seq) const
 {
     if (!full_)
         return {};
-    std::shared_lock{mtx_};
+    std::shared_lock lck{mtx_};
     successorReqCounter_++;
     if (seq != latestSeq_)
         return {};
@@ -146,7 +146,7 @@ LedgerCache::getObjectHitRate() const
 {
     if (!objectReqCounter_)
         return 1;
-    return ((float)objectHitCounter_) / objectReqCounter_;
+    return static_cast<float>(objectHitCounter_) / objectReqCounter_;
 }
 
 float
@@ -154,7 +154,7 @@ LedgerCache::getSuccessorHitRate() const
 {
     if (!successorReqCounter_)
         return 1;
-    return ((float)successorHitCounter_) / successorReqCounter_;
+    return static_cast<float>(successorHitCounter_) / successorReqCounter_;
 }
 
-}  // namespace Backend
+}  // namespace data

src/data/LedgerCache.h

@@ -21,15 +21,18 @@
 
 #include <ripple/basics/base_uint.h>
 #include <ripple/basics/hardened_hash.h>
-#include <backend/Types.h>
+#include <data/Types.h>
 #include <map>
 #include <mutex>
 #include <shared_mutex>
 #include <utility>
 #include <vector>
 
-namespace Backend {
+namespace data {
 
+/**
+ * @brief Cache for an entire ledger.
+ */
 class LedgerCache
 {
     struct CacheEntry
@@ -57,42 +60,95 @@ class LedgerCache
     std::unordered_set<ripple::uint256, ripple::hardened_hash<>> deletes_;
 
 public:
-    // Update the cache with new ledger objects set isBackground to true when writing old data from a background thread
+    /**
+     * @brief Update the cache with new ledger objects.
+     *
+     * @param blobs The ledger objects to update cache with
+     * @param seq The sequence to update cache for
+     * @param isBackground Should be set to true when writing old data from a background thread
+     */
     void
     update(std::vector<LedgerObject> const& blobs, uint32_t seq, bool isBackground = false);
 
+    /**
+     * @brief Fetch a cached object by its key and sequence number.
+     *
+     * @param key The key to fetch for
+     * @param seq The sequence to fetch for
+     * @return If found in cache, will return the cached Blob; otherwise nullopt is returned
+     */
     std::optional<Blob>
     get(ripple::uint256 const& key, uint32_t seq) const;
 
-    // always returns empty optional if isFull() is false
+    /**
+     * @brief Gets a cached successor.
+     *
+     * Note: This function always returns std::nullopt when @ref isFull() returns false.
+     *
+     * @param key The key to fetch for
+     * @param seq The sequence to fetch for
+     * @return If found in cache, will return the cached successor; otherwise nullopt is returned
+     */
     std::optional<LedgerObject>
     getSuccessor(ripple::uint256 const& key, uint32_t seq) const;
 
-    // always returns empty optional if isFull() is false
+    /**
+     * @brief Gets a cached predcessor.
+     *
+     * Note: This function always returns std::nullopt when @ref isFull() returns false.
+     *
+     * @param key The key to fetch for
+     * @param seq The sequence to fetch for
+     * @return If found in cache, will return the cached predcessor; otherwise nullopt is returned
+     */
     std::optional<LedgerObject>
     getPredecessor(ripple::uint256 const& key, uint32_t seq) const;
 
+    /**
+     * @brief Disables the cache.
+     */
     void
     setDisabled();
 
+    /**
+     * @brief Sets the full flag to true.
+     *
+     * This is used when cache loaded in its entirety at startup of the application. This can be either loaded from DB,
+     * populated together with initial ledger download (on first run) or downloaded from a peer node (specified in
+     * config).
+     */
     void
     setFull();
 
+    /**
+     * @return The latest ledger sequence for which cache is available.
+     */
     uint32_t
     latestLedgerSequence() const;
 
-    // whether the cache has all data for the most recent ledger
+    /**
+     * @return true if the cache has all data for the most recent ledger; false otherwise
+     */
     bool
     isFull() const;
 
+    /**
+     * @return The total size of the cache.
+     */
     size_t
     size() const;
 
+    /**
+     * @return A number representing the success rate of hitting an object in the cache versus missing it.
+     */
     float
     getObjectHitRate() const;
 
+    /**
+     * @return A number representing the success rate of hitting a successor in the cache versus missing it.
+     */
     float
     getSuccessorHitRate() const;
 };
 
-}  // namespace Backend
+}  // namespace data

src/data/README.md

@@ -1,4 +1,5 @@
-# Clio Backend
+# Backend
+
 ## Background
 The backend of Clio is responsible for handling the proper reading and writing of past ledger data from and to a given database. As of right now, Cassandra and ScyllaDB are the only supported databases that are production-ready. Support for database types can be easily extended by creating new implementations which implements the virtual methods of `BackendInterface.h`. Then, use the Factory Object Design Pattern to simply add logic statements to `BackendFactory.h` that return the new database interface for a specific `type` in Clio's configuration file. 
 

src/data/Types.h

@@ -21,20 +21,23 @@
 
 #include <ripple/basics/base_uint.h>
 #include <ripple/protocol/AccountID.h>
+
 #include <optional>
 #include <string>
 #include <vector>
 
-namespace Backend {
-
-// *** return types
+namespace data {
 
 using Blob = std::vector<unsigned char>;
 
+/**
+ * @brief Represents an object in the ledger.
+ */
 struct LedgerObject
 {
     ripple::uint256 key;
     Blob blob;
+
     bool
     operator==(const LedgerObject& other) const
     {
@@ -42,16 +45,27 @@ struct LedgerObject
     }
 };
 
+/**
+ * @brief Represents a page of LedgerObjects.
+ */
 struct LedgerPage
 {
     std::vector<LedgerObject> objects;
     std::optional<ripple::uint256> cursor;
 };
+
+/**
+ * @brief Represents a page of book offer objects.
+ */
 struct BookOffersPage
 {
     std::vector<LedgerObject> offers;
     std::optional<ripple::uint256> cursor;
 };
+
+/**
+ * @brief Represents a transaction and its metadata bundled together.
+ */
 struct TransactionAndMetadata
 {
     Blob transaction;
@@ -85,10 +99,13 @@ struct TransactionAndMetadata
     }
 };
 
+/**
+ * @brief Represents a cursor into the transactions table.
+ */
 struct TransactionsCursor
 {
-    std::uint32_t ledgerSequence;
-    std::uint32_t transactionIndex;
+    std::uint32_t ledgerSequence = 0;
+    std::uint32_t transactionIndex = 0;
 
     TransactionsCursor() = default;
     TransactionsCursor(std::uint32_t ledgerSequence, std::uint32_t transactionIndex)
@@ -101,6 +118,8 @@ struct TransactionsCursor
     {
     }
 
+    TransactionsCursor(TransactionsCursor const&) = default;
+
     TransactionsCursor&
     operator=(TransactionsCursor const&) = default;
 
@@ -114,12 +133,18 @@ struct TransactionsCursor
     }
 };
 
+/**
+ * @brief Represests a bundle of transactions with metadata and a cursor to the next page.
+ */
 struct TransactionsAndCursor
 {
     std::vector<TransactionAndMetadata> txns;
     std::optional<TransactionsCursor> cursor;
 };
 
+/**
+ * @brief Represents a NFToken.
+ */
 struct NFT
 {
     ripple::uint256 tokenID;
@@ -143,9 +168,8 @@ struct NFT
     {
     }
 
-    // clearly two tokens are the same if they have the same ID, but this
-    // struct stores the state of a given token at a given ledger sequence, so
-    // we also need to compare with ledgerSequence
+    // clearly two tokens are the same if they have the same ID, but this struct stores the state of a given token at a
+    // given ledger sequence, so we also need to compare with ledgerSequence.
     bool
     operator==(NFT const& other) const
     {
@@ -153,12 +177,17 @@ struct NFT
     }
 };
 
+/**
+ * @brief Stores a range of sequences as a min and max pair.
+ */
 struct LedgerRange
 {
-    std::uint32_t minSequence;
-    std::uint32_t maxSequence;
+    std::uint32_t minSequence = 0;
+    std::uint32_t maxSequence = 0;
 };
+
 constexpr ripple::uint256 firstKey{"0000000000000000000000000000000000000000000000000000000000000000"};
 constexpr ripple::uint256 lastKey{"FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF"};
 constexpr ripple::uint256 hi192{"0000000000000000000000000000000000000000000000001111111111111111"};
-}  // namespace Backend
+
+}  // namespace data

src/data/cassandra/Concepts.h

@@ -19,7 +19,7 @@
 
 #pragma once
 
-#include <backend/cassandra/Types.h>
+#include <data/cassandra/Types.h>
 
 #include <boost/asio/spawn.hpp>
 
@@ -28,8 +28,11 @@
 #include <optional>
 #include <string>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
+/**
+ * @brief The requirements of a settings provider.
+ */
 // clang-format off
 template <typename T>
 concept SomeSettingsProvider = requires(T a) {
@@ -41,6 +44,9 @@ concept SomeSettingsProvider = requires(T a) {
 };
 // clang-format on
 
+/**
+ * @brief The requirements of an execution strategy.
+ */
 // clang-format off
 template <typename T>
 concept SomeExecutionStrategy = requires(
@@ -66,6 +72,9 @@ concept SomeExecutionStrategy = requires(
 };
 // clang-format on
 
+/**
+ * @brief The requirements of a retry policy.
+ */
 // clang-format off
 template <typename T>
 concept SomeRetryPolicy = requires(T a, boost::asio::io_context ioc, CassandraError err, uint32_t attempt) {
@@ -76,4 +85,4 @@ concept SomeRetryPolicy = requires(T a, boost::asio::io_context ioc, CassandraEr
 };
 // clang-format on
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/Error.h

@@ -23,10 +23,10 @@
 
 #include <string>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 /**
- * @brief A simple container for both error message and error code
+ * @brief A simple container for both error message and error code.
  */
 class CassandraError
 {
@@ -67,18 +67,27 @@ public:
         return os;
     }
 
+    /**
+     * @return The final error message as a std::string
+     */
     std::string
     message() const
     {
         return message_;
     }
 
+    /**
+     * @return The error code
+     */
     uint32_t
     code() const
     {
         return code_;
     }
 
+    /**
+     * @return true if the wrapped error is considered a timeout; false otherwise
+     */
     bool
     isTimeout() const
     {
@@ -89,6 +98,9 @@ public:
         return false;
     }
 
+    /**
+     * @return true if the wrapped error is an invalid query; false otherwise
+     */
     bool
     isInvalidQuery() const
     {
@@ -96,4 +108,4 @@ public:
     }
 };
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/Handle.cpp

@@ -17,9 +17,9 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/Handle.h>
+#include <data/cassandra/Handle.h>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 Handle::Handle(Settings clusterSettings) : cluster_{clusterSettings}
 {
@@ -152,4 +152,4 @@ Handle::prepare(std::string_view query) const
         throw std::runtime_error(rc.error().message());
 }
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/Handle.h

@@ -19,15 +19,15 @@
 
 #pragma once
 
-#include <backend/cassandra/Error.h>
-#include <backend/cassandra/Types.h>
-#include <backend/cassandra/impl/Batch.h>
-#include <backend/cassandra/impl/Cluster.h>
-#include <backend/cassandra/impl/Future.h>
-#include <backend/cassandra/impl/ManagedObject.h>
-#include <backend/cassandra/impl/Result.h>
-#include <backend/cassandra/impl/Session.h>
-#include <backend/cassandra/impl/Statement.h>
+#include <data/cassandra/Error.h>
+#include <data/cassandra/Types.h>
+#include <data/cassandra/impl/Batch.h>
+#include <data/cassandra/impl/Cluster.h>
+#include <data/cassandra/impl/Future.h>
+#include <data/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/impl/Result.h>
+#include <data/cassandra/impl/Session.h>
+#include <data/cassandra/impl/Statement.h>
 #include <util/Expected.h>
 
 #include <cassandra.h>
@@ -37,7 +37,7 @@
 #include <iterator>
 #include <vector>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 /**
  * @brief Represents a handle to the cassandra database cluster
@@ -57,28 +57,31 @@ public:
     using ResultType = Result;
 
     /**
-     * @brief Construct a new handle from a @ref Settings object
+     * @brief Construct a new handle from a @ref detail::Settings object.
+     *
+     * @param clusterSettings The settings to use
      */
     explicit Handle(Settings clusterSettings = Settings::defaultSettings());
 
     /**
-     * @brief Construct a new handle with default settings and only by setting
-     * the contact points
+     * @brief Construct a new handle with default settings and only by setting the contact points.
+     *
+     * @param contactPoints The contact points to use instead of settings
      */
     explicit Handle(std::string_view contactPoints);
 
     /**
-     * @brief Disconnects gracefully if possible
+     * @brief Disconnects gracefully if possible.
      */
     ~Handle();
 
     /**
-     * @brief Move is supported
+     * @brief Move is supported.
      */
     Handle(Handle&&) = default;
 
     /**
-     * @brief Connect to the cluster asynchronously
+     * @brief Connect to the cluster asynchronously.
      *
      * @return A future
      */
@@ -86,31 +89,37 @@ public:
     asyncConnect() const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
      * See @ref asyncConnect() const for how this works.
+     *
+     * @return Possibly an error
      */
     [[nodiscard]] MaybeErrorType
     connect() const;
 
     /**
-     * @brief Connect to the the specified keyspace asynchronously
+     * @brief Connect to the the specified keyspace asynchronously.
      *
+     * @param keyspace The keyspace to use
      * @return A future
      */
     [[nodiscard]] FutureType
     asyncConnect(std::string_view keyspace) const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
      * See @ref asyncConnect(std::string_view) const for how this works.
+     *
+     * @param keyspace The keyspace to use
+     * @return Possibly an error
      */
     [[nodiscard]] MaybeErrorType
     connect(std::string_view keyspace) const;
 
     /**
-     * @brief Disconnect from the cluster asynchronously
+     * @brief Disconnect from the cluster asynchronously.
      *
      * @return A future
      */
@@ -118,32 +127,40 @@ public:
     asyncDisconnect() const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
      * See @ref asyncDisconnect() const for how this works.
+     *
+     * @return Possibly an error
      */
     [[maybe_unused]] MaybeErrorType
     disconnect() const;
 
     /**
-     * @brief Reconnect to the the specified keyspace asynchronously
+     * @brief Reconnect to the the specified keyspace asynchronously.
      *
+     * @param keyspace The keyspace to use
      * @return A future
      */
     [[nodiscard]] FutureType
     asyncReconnect(std::string_view keyspace) const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
      * See @ref asyncReconnect(std::string_view) const for how this works.
+     *
+     * @param keyspace The keyspace to use
+     * @return Possibly an error
      */
     [[nodiscard]] MaybeErrorType
     reconnect(std::string_view keyspace) const;
 
     /**
-     * @brief Execute a simple query with optional args asynchronously
+     * @brief Execute a simple query with optional args asynchronously.
      *
+     * @param query The query to execute
+     * @param args The arguments to bind for execution
      * @return A future
      */
     template <typename... Args>
@@ -155,10 +172,13 @@ public:
     }
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
-     * See @ref asyncExecute(std::string_view, Args&&...) const for how this
-     * works.
+     * See asyncExecute(std::string_view, Args&&...) const for how this works.
+     *
+     * @param query The query to execute
+     * @param args The arguments to bind for execution
+     * @return The result or an error
      */
     template <typename... Args>
     [[maybe_unused]] ResultOrErrorType
@@ -168,30 +188,34 @@ public:
     }
 
     /**
-     * @brief Execute each of the statements asynchronously
+     * @brief Execute each of the statements asynchronously.
      *
-     * Batched version is not always the right option. Especially since it only
-     * supports INSERT, UPDATE and DELETE statements.
-     * This can be used as an alternative when statements need to execute in
-     * bulk.
+     * Batched version is not always the right option.
+     * Especially since it only supports INSERT, UPDATE and DELETE statements.
+     * This can be used as an alternative when statements need to execute in bulk.
      *
+     * @param statements The statements to execute
      * @return A vector of future objects
      */
     [[nodiscard]] std::vector<FutureType>
     asyncExecuteEach(std::vector<StatementType> const& statements) const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
-     * See @ref asyncExecuteEach(std::vector<StatementType> const&) const for
-     * how this works.
+     * See @ref asyncExecuteEach(std::vector<StatementType> const&) const for how this works.
+     *
+     * @param statements The statements to execute
+     * @return Possibly an error
      */
     [[maybe_unused]] MaybeErrorType
     executeEach(std::vector<StatementType> const& statements) const;
 
     /**
-     * @brief Execute a prepared statement with optional args asynchronously
+     * @brief Execute a prepared statement with optional args asynchronously.
      *
+     * @param statement The prepared statement to execute
+     * @param args The arguments to bind for execution
      * @return A future
      */
     template <typename... Args>
@@ -203,10 +227,13 @@ public:
     }
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
-     * See @ref asyncExecute(std::vector<StatementType> const&, Args&&...) const
-     * for how this works.
+     * See asyncExecute(std::vector<StatementType> const&, Args&&...) const for how this works.
+     *
+     * @param statement The prepared statement to bind and execute
+     * @param args The arguments to bind for execution
+     * @return The result or an error
      */
     template <typename... Args>
     [[maybe_unused]] ResultOrErrorType
@@ -216,61 +243,70 @@ public:
     }
 
     /**
-     * @brief Execute one (bound or simple) statements asynchronously
+     * @brief Execute one (bound or simple) statements asynchronously.
      *
+     * @param statement The statement to execute
      * @return A future
      */
     [[nodiscard]] FutureType
     asyncExecute(StatementType const& statement) const;
 
     /**
-     * @brief Execute one (bound or simple) statements asynchronously with a
-     * callback
+     * @brief Execute one (bound or simple) statements asynchronously with a callback.
      *
+     * @param statement The statement to execute
+     * @param cb The callback to execute when data is ready
      * @return A future that holds onto the callback provided
      */
     [[nodiscard]] FutureWithCallbackType
     asyncExecute(StatementType const& statement, std::function<void(ResultOrErrorType)>&& cb) const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
-     * See @ref asyncExecute(StatementType const&) const for how this
-     * works.
+     * See @ref asyncExecute(StatementType const&) const for how this works.
+     *
+     * @param statement The statement to execute
+     * @return The result or an error
      */
     [[maybe_unused]] ResultOrErrorType
     execute(StatementType const& statement) const;
 
     /**
-     * @brief Execute a batch of (bound or simple) statements asynchronously
+     * @brief Execute a batch of (bound or simple) statements asynchronously.
      *
+     * @param statements The statements to execute
      * @return A future
      */
     [[nodiscard]] FutureType
     asyncExecute(std::vector<StatementType> const& statements) const;
 
     /**
-     * @brief Synchonous version of the above
+     * @brief Synchonous version of the above.
      *
-     * See @ref asyncExecute(std::vector<StatementType> const&) const for how
-     * this works.
+     * See @ref asyncExecute(std::vector<StatementType> const&) const for how this works.
+     *
+     * @param statements The statements to execute
+     * @return Possibly an error
      */
     [[maybe_unused]] MaybeErrorType
     execute(std::vector<StatementType> const& statements) const;
 
     /**
-     * @brief Execute a batch of (bound or simple) statements asynchronously
-     * with a completion callback
+     * @brief Execute a batch of (bound or simple) statements asynchronously with a completion callback.
      *
+     * @param statements The statements to execute
+     * @param cb The callback to execute when data is ready
      * @return A future that holds onto the callback provided
      */
     [[nodiscard]] FutureWithCallbackType
     asyncExecute(std::vector<StatementType> const& statements, std::function<void(ResultOrErrorType)>&& cb) const;
 
     /**
-     * @brief Prepare a statement
+     * @brief Prepare a statement.
      *
-     * @return A @ref PreparedStatementType
+     * @param query
+     * @return A prepared statement
      * @throws std::runtime_error with underlying error description on failure
      */
     [[nodiscard]] PreparedStatementType
@@ -278,12 +314,13 @@ public:
 };
 
 /**
- * @brief Extracts the results into series of std::tuple<Types...> by creating a
- * simple wrapper with an STL input iterator inside.
+ * @brief Extracts the results into series of std::tuple<Types...> by creating a simple wrapper with an STL input
+ * iterator inside.
  *
  * You can call .begin() and .end() in order to iterate as usual.
- * This also means that you can use it in a range-based for or with some
- * algorithms.
+ * This also means that you can use it in a range-based for or with some algorithms.
+ *
+ * @param result The result to iterate
  */
 template <typename... Types>
 [[nodiscard]] detail::ResultExtractor<Types...>
@@ -292,4 +329,4 @@ extract(Handle::ResultType const& result)
     return {result};
 }
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/Schema.h

@@ -19,17 +19,17 @@
 
 #pragma once
 
-#include <backend/cassandra/Concepts.h>
-#include <backend/cassandra/Handle.h>
-#include <backend/cassandra/SettingsProvider.h>
-#include <backend/cassandra/Types.h>
-#include <config/Config.h>
-#include <log/Logger.h>
+#include <data/cassandra/Concepts.h>
+#include <data/cassandra/Handle.h>
+#include <data/cassandra/SettingsProvider.h>
+#include <data/cassandra/Types.h>
 #include <util/Expected.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
 
 #include <fmt/compile.h>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 template <SomeSettingsProvider SettingsProviderType>
 [[nodiscard]] std::string inline qualifiedTableName(SettingsProviderType const& provider, std::string_view name)
@@ -38,17 +38,12 @@ template <SomeSettingsProvider SettingsProviderType>
 }
 
 /**
- * @brief Manages the DB schema and provides access to prepared statements
+ * @brief Manages the DB schema and provides access to prepared statements.
  */
 template <SomeSettingsProvider SettingsProviderType>
 class Schema
 {
-    // Current schema version.
-    // Update this everytime you update the schema.
-    // Migrations will be ran automatically based on this value.
-    static constexpr uint16_t version = 1u;
-
-    clio::Logger log_{"Backend"};
+    util::Logger log_{"Backend"};
     std::reference_wrapper<SettingsProviderType const> settingsProvider_;
 
 public:
@@ -261,7 +256,7 @@ public:
     }();
 
     /**
-     * @brief Prepared statements holder
+     * @brief Prepared statements holder.
      */
     class Statements
     {
@@ -528,21 +523,19 @@ public:
                 SELECT hash, seq_idx 
                   FROM {}               
                  WHERE account = ?
-                   AND seq_idx <= ?
+                   AND seq_idx < ?
                  LIMIT ?
                 )",
                 qualifiedTableName(settingsProvider_.get(), "account_tx")));
         }();
 
-        // the smallest transaction idx is 0, we use uint to store the transaction index, so we shall use ">=" to
-        // include it(the transaction with idx 0) in the result
         PreparedStatement selectAccountTxForward = [this]() {
             return handle_.get().prepare(fmt::format(
                 R"(
                 SELECT hash, seq_idx 
                   FROM {}               
                  WHERE account = ?
-                   AND seq_idx >= ?
+                   AND seq_idx > ?
               ORDER BY seq_idx ASC 
                  LIMIT ?
                 )",
@@ -643,18 +636,18 @@ public:
     };
 
     /**
-     * @brief Recreates the prepared statements
+     * @brief Recreates the prepared statements.
      */
     void
     prepareStatements(Handle const& handle)
     {
-        log_.info() << "Preparing cassandra statements";
+        LOG(log_.info()) << "Preparing cassandra statements";
         statements_ = std::make_unique<Statements>(settingsProvider_, handle);
-        log_.info() << "Finished preparing statements";
+        LOG(log_.info()) << "Finished preparing statements";
     }
 
     /**
-     * @brief Provides access to statements
+     * @brief Provides access to statements.
      */
     std::unique_ptr<Statements> const&
     operator->() const
@@ -666,4 +659,4 @@ private:
     std::unique_ptr<Statements> statements_{nullptr};
 };
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/SettingsProvider.cpp

@@ -17,17 +17,18 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/SettingsProvider.h>
-#include <backend/cassandra/impl/Cluster.h>
-#include <backend/cassandra/impl/Statement.h>
-#include <config/Config.h>
+#include <data/cassandra/SettingsProvider.h>
+#include <data/cassandra/impl/Cluster.h>
+#include <data/cassandra/impl/Statement.h>
+#include <util/config/Config.h>
 
 #include <boost/json.hpp>
 
+#include <fstream>
 #include <string>
 #include <thread>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 namespace detail {
 inline Settings::ContactPoints
@@ -38,7 +39,7 @@ tag_invoke(boost::json::value_to_tag<Settings::ContactPoints>, boost::json::valu
             "Feed entire Cassandra section to parse "
             "Settings::ContactPoints instead");
 
-    clio::Config obj{value};
+    util::Config obj{value};
     Settings::ContactPoints out;
 
     out.contactPoints = obj.valueOrThrow<std::string>("contact_points", "`contact_points` must be a string");
@@ -56,7 +57,7 @@ tag_invoke(boost::json::value_to_tag<Settings::SecureConnectionBundle>, boost::j
 }
 }  // namespace detail
 
-SettingsProvider::SettingsProvider(clio::Config const& cfg, uint16_t ttl)
+SettingsProvider::SettingsProvider(util::Config const& cfg, uint16_t ttl)
     : config_{cfg}
     , keyspace_{cfg.valueOr<std::string>("keyspace", "clio")}
     , tablePrefix_{cfg.maybeValue<std::string>("table_prefix")}
@@ -115,6 +116,19 @@ SettingsProvider::parseSettings() const
         config_.valueOr<uint32_t>("max_write_requests_outstanding", settings.maxWriteRequestsOutstanding);
     settings.maxReadRequestsOutstanding =
         config_.valueOr<uint32_t>("max_read_requests_outstanding", settings.maxReadRequestsOutstanding);
+    settings.coreConnectionsPerHost =
+        config_.valueOr<uint32_t>("core_connections_per_host", settings.coreConnectionsPerHost);
+
+    settings.queueSizeIO = config_.maybeValue<uint32_t>("queue_size_io");
+
+    auto const connectTimeoutSecond = config_.maybeValue<uint32_t>("connect_timeout");
+    if (connectTimeoutSecond)
+        settings.connectionTimeout = std::chrono::milliseconds{*connectTimeoutSecond * 1000};
+
+    auto const requestTimeoutSecond = config_.maybeValue<uint32_t>("request_timeout");
+    if (requestTimeoutSecond)
+        settings.requestTimeout = std::chrono::milliseconds{*requestTimeoutSecond * 1000};
+
     settings.certificate = parseOptionalCertificate();
     settings.username = config_.maybeValue<std::string>("username");
     settings.password = config_.maybeValue<std::string>("password");
@@ -122,4 +136,4 @@ SettingsProvider::parseSettings() const
     return settings;
 }
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/SettingsProvider.h

@@ -19,20 +19,20 @@
 
 #pragma once
 
-#include <backend/cassandra/Handle.h>
-#include <backend/cassandra/Types.h>
-#include <config/Config.h>
-#include <log/Logger.h>
+#include <data/cassandra/Handle.h>
+#include <data/cassandra/Types.h>
 #include <util/Expected.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 /**
- * @brief Provides settings for @ref CassandraBackend
+ * @brief Provides settings for @ref BasicCassandraBackend.
  */
 class SettingsProvider
 {
-    clio::Config config_;
+    util::Config config_;
 
     std::string keyspace_;
     std::optional<std::string> tablePrefix_;
@@ -41,34 +41,50 @@ class SettingsProvider
     Settings settings_;
 
 public:
-    explicit SettingsProvider(clio::Config const& cfg, uint16_t ttl = 0);
+    /**
+     * @brief Create a settings provider from the specified config.
+     *
+     * @param cfg The config of Clio to use
+     * @param ttl Time to live setting
+     */
+    explicit SettingsProvider(util::Config const& cfg, uint16_t ttl = 0);
 
-    /*! Get the cluster settings */
+    /**
+     * @return The cluster settings
+     */
     [[nodiscard]] Settings
     getSettings() const;
 
-    /*! Get the specified keyspace */
+    /**
+     * @return The specified keyspace
+     */
     [[nodiscard]] inline std::string
     getKeyspace() const
     {
         return keyspace_;
     }
 
-    /*! Get an optional table prefix to use in all queries */
+    /**
+     * @return The optional table prefix to use in all queries
+     */
     [[nodiscard]] inline std::optional<std::string>
     getTablePrefix() const
     {
         return tablePrefix_;
     }
 
-    /*! Get the replication factor */
+    /**
+     * @return The replication factor
+     */
     [[nodiscard]] inline uint16_t
     getReplicationFactor() const
     {
         return replicationFactor_;
     }
 
-    /*! Get the default time to live to use in all `create` queries */
+    /**
+     * @return The default time to live to use in all `create` queries
+     */
     [[nodiscard]] inline uint16_t
     getTtl() const
     {
@@ -83,4 +99,4 @@ private:
     parseSettings() const;
 };
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/Types.h

@@ -23,7 +23,7 @@
 
 #include <string>
 
-namespace Backend::Cassandra {
+namespace data::cassandra {
 
 namespace detail {
 struct Settings;
@@ -64,4 +64,4 @@ using MaybeError = util::Expected<void, CassandraError>;
 using ResultOrError = util::Expected<Result, CassandraError>;
 using Error = util::Unexpected<CassandraError>;
 
-}  // namespace Backend::Cassandra
+}  // namespace data::cassandra

src/data/cassandra/impl/AsyncExecutor.h

@@ -19,19 +19,19 @@
 
 #pragma once
 
-#include <backend/cassandra/Concepts.h>
-#include <backend/cassandra/Handle.h>
-#include <backend/cassandra/Types.h>
-#include <backend/cassandra/impl/RetryPolicy.h>
-#include <log/Logger.h>
+#include <data/cassandra/Concepts.h>
+#include <data/cassandra/Handle.h>
+#include <data/cassandra/Types.h>
+#include <data/cassandra/impl/RetryPolicy.h>
 #include <util/Expected.h>
+#include <util/log/Logger.h>
 
 #include <boost/asio.hpp>
 
 #include <functional>
 #include <memory>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 /**
  * @brief A query executor with a changable retry policy
@@ -53,7 +53,7 @@ class AsyncExecutor : public std::enable_shared_from_this<AsyncExecutor<Statemen
     using FutureWithCallbackType = typename HandleType::FutureWithCallbackType;
     using CallbackType = std::function<void(typename HandleType::ResultOrErrorType)>;
 
-    clio::Logger log_{"Backend"};
+    util::Logger log_{"Backend"};
 
     StatementType data_;
     RetryPolicyType retryPolicy_;
@@ -116,4 +116,4 @@ private:
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Batch.cpp

@@ -17,9 +17,9 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/Error.h>
-#include <backend/cassandra/impl/Batch.h>
-#include <backend/cassandra/impl/Statement.h>
+#include <data/cassandra/Error.h>
+#include <data/cassandra/impl/Batch.h>
+#include <data/cassandra/impl/Statement.h>
 #include <util/Expected.h>
 
 #include <exception>
@@ -29,10 +29,9 @@ namespace {
 static constexpr auto batchDeleter = [](CassBatch* ptr) { cass_batch_free(ptr); };
 };
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
-// todo: use an appropritae value instead of CASS_BATCH_TYPE_LOGGED for
-// different use cases
+// TODO: Use an appropriate value instead of CASS_BATCH_TYPE_LOGGED for different use cases
 Batch::Batch(std::vector<Statement> const& statements)
     : ManagedObject{cass_batch_new(CASS_BATCH_TYPE_LOGGED), batchDeleter}
 {
@@ -53,4 +52,4 @@ Batch::add(Statement const& statement)
     return {};
 }
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Batch.h

@@ -19,12 +19,12 @@
 
 #pragma once
 
-#include <backend/cassandra/Types.h>
-#include <backend/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/Types.h>
+#include <data/cassandra/impl/ManagedObject.h>
 
 #include <cassandra.h>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 struct Batch : public ManagedObject<CassBatch>
 {
@@ -34,4 +34,4 @@ struct Batch : public ManagedObject<CassBatch>
     add(Statement const& statement);
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Cluster.cpp

@@ -17,11 +17,13 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/impl/Cluster.h>
-#include <backend/cassandra/impl/SslContext.h>
-#include <backend/cassandra/impl/Statement.h>
+#include <data/cassandra/impl/Cluster.h>
+#include <data/cassandra/impl/SslContext.h>
+#include <data/cassandra/impl/Statement.h>
 #include <util/Expected.h>
 
+#include <fmt/core.h>
+
 #include <exception>
 #include <vector>
 
@@ -39,7 +41,7 @@ template <class... Ts>
 overloadSet(Ts...) -> overloadSet<Ts...>;
 };  // namespace
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 Cluster::Cluster(Settings const& settings) : ManagedObject{cass_cluster_new(), clusterDeleter}
 {
@@ -48,44 +50,40 @@ Cluster::Cluster(Settings const& settings) : ManagedObject{cass_cluster_new(), c
     cass_cluster_set_token_aware_routing(*this, cass_true);
     if (auto const rc = cass_cluster_set_protocol_version(*this, CASS_PROTOCOL_VERSION_V4); rc != CASS_OK)
     {
-        throw std::runtime_error(std::string{"Error setting cassandra protocol version to v4: "} + cass_error_desc(rc));
+        throw std::runtime_error(
+            fmt::format("Error setting cassandra protocol version to v4: {}", cass_error_desc(rc)));
     }
 
     if (auto const rc = cass_cluster_set_num_threads_io(*this, settings.threads); rc != CASS_OK)
     {
         throw std::runtime_error(
-            std::string{"Error setting cassandra io threads to "} + to_string(settings.threads) + ": " +
-            cass_error_desc(rc));
+            fmt::format("Error setting cassandra io threads to {}: {}", settings.threads, cass_error_desc(rc)));
     }
 
     cass_log_set_level(settings.enableLog ? CASS_LOG_TRACE : CASS_LOG_DISABLED);
     cass_cluster_set_connect_timeout(*this, settings.connectionTimeout.count());
     cass_cluster_set_request_timeout(*this, settings.requestTimeout.count());
 
-    // TODO: other options to experiment with and consider later:
-    // cass_cluster_set_max_concurrent_requests_threshold(*this, 10000);
-    // cass_cluster_set_queue_size_event(*this, 100000);
-    // cass_cluster_set_queue_size_io(*this, 100000);
-    // cass_cluster_set_write_bytes_high_water_mark(*this, 16 * 1024 * 1024);  // 16mb
-    // cass_cluster_set_write_bytes_low_water_mark(*this, 8 * 1024 * 1024);  // half of allowance
-    // cass_cluster_set_pending_requests_high_water_mark(*this, 5000);
-    // cass_cluster_set_pending_requests_low_water_mark(*this, 2500);  // half
-    // cass_cluster_set_max_requests_per_flush(*this, 1000);
-    // cass_cluster_set_max_concurrent_creation(*this, 8);
-    // cass_cluster_set_max_connections_per_host(*this, 6);
-    // cass_cluster_set_core_connections_per_host(*this, 4);
-    // cass_cluster_set_constant_speculative_execution_policy(*this, 1000, 1024);
-
-    if (auto const rc = cass_cluster_set_queue_size_io(
-            *this, settings.maxWriteRequestsOutstanding + settings.maxReadRequestsOutstanding);
+    if (auto const rc = cass_cluster_set_core_connections_per_host(*this, settings.coreConnectionsPerHost);
         rc != CASS_OK)
     {
-        throw std::runtime_error(std::string{"Could not set queue size for IO per host: "} + cass_error_desc(rc));
+        throw std::runtime_error(fmt::format("Could not set core connections per host: {}", cass_error_desc(rc)));
+    }
+
+    auto const queueSize =
+        settings.queueSizeIO.value_or(settings.maxWriteRequestsOutstanding + settings.maxReadRequestsOutstanding);
+    if (auto const rc = cass_cluster_set_queue_size_io(*this, queueSize); rc != CASS_OK)
+    {
+        throw std::runtime_error(fmt::format("Could not set queue size for IO per host: {}", cass_error_desc(rc)));
     }
 
     setupConnection(settings);
     setupCertificate(settings);
     setupCredentials(settings);
+
+    LOG(log_.info()) << "Threads: " << settings.threads;
+    LOG(log_.info()) << "Core connections per host: " << settings.coreConnectionsPerHost;
+    LOG(log_.info()) << "IO queue size: " << queueSize;
 }
 
 void
@@ -104,11 +102,12 @@ Cluster::setupContactPoints(Settings::ContactPoints const& points)
     using std::to_string;
     auto throwErrorIfNeeded = [](CassError rc, std::string const& label, std::string const& value) {
         if (rc != CASS_OK)
-            throw std::runtime_error("Cassandra: Error setting " + label + " [" + value + "]: " + cass_error_desc(rc));
+            throw std::runtime_error(
+                fmt::format("Cassandra: Error setting {} [{}]: {}", label, value, cass_error_desc(rc)));
     };
 
     {
-        log_.debug() << "Attempt connection using contact points: " << points.contactPoints;
+        LOG(log_.debug()) << "Attempt connection using contact points: " << points.contactPoints;
         auto const rc = cass_cluster_set_contact_points(*this, points.contactPoints.data());
         throwErrorIfNeeded(rc, "contact_points", points.contactPoints);
     }
@@ -123,10 +122,10 @@ Cluster::setupContactPoints(Settings::ContactPoints const& points)
 void
 Cluster::setupSecureBundle(Settings::SecureConnectionBundle const& bundle)
 {
-    log_.debug() << "Attempt connection using secure bundle";
+    LOG(log_.debug()) << "Attempt connection using secure bundle";
     if (auto const rc = cass_cluster_set_cloud_secure_connection_bundle(*this, bundle.bundle.data()); rc != CASS_OK)
     {
-        throw std::runtime_error("Failed to connect using secure connection bundle" + bundle.bundle);
+        throw std::runtime_error("Failed to connect using secure connection bundle " + bundle.bundle);
     }
 }
 
@@ -136,7 +135,7 @@ Cluster::setupCertificate(Settings const& settings)
     if (not settings.certificate)
         return;
 
-    log_.debug() << "Configure SSL context";
+    LOG(log_.debug()) << "Configure SSL context";
     SslContext context = SslContext(*settings.certificate);
     cass_cluster_set_ssl(*this, context);
 }
@@ -147,8 +146,8 @@ Cluster::setupCredentials(Settings const& settings)
     if (not settings.username || not settings.password)
         return;
 
-    log_.debug() << "Set credentials; username: " << settings.username.value();
+    LOG(log_.debug()) << "Set credentials; username: " << settings.username.value();
     cass_cluster_set_credentials(*this, settings.username.value().c_str(), settings.password.value().c_str());
 }
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Cluster.h

@@ -19,8 +19,8 @@
 
 #pragma once
 
-#include <backend/cassandra/impl/ManagedObject.h>
-#include <log/Logger.h>
+#include <data/cassandra/impl/ManagedObject.h>
+#include <util/log/Logger.h>
 
 #include <cassandra.h>
 
@@ -31,32 +31,71 @@
 #include <thread>
 #include <variant>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
+// TODO: move Settings to public interface, not detail
+
+/**
+ * @brief Bundles all cassandra settings in one place.
+ */
 struct Settings
 {
+    /**
+     * @brief Represents the configuration of contact points for cassandra.
+     */
     struct ContactPoints
     {
         std::string contactPoints = "127.0.0.1";  // defaults to localhost
-        std::optional<uint16_t> port;
+        std::optional<uint16_t> port = {};
     };
 
+    /**
+     * @brief Represents the configuration of a secure connection bundle.
+     */
     struct SecureConnectionBundle
     {
         std::string bundle;  // no meaningful default
     };
 
+    /** @brief Enables or disables cassandra driver logger */
     bool enableLog = false;
+
+    /** @brief Connect timeout specified in milliseconds */
     std::chrono::milliseconds connectionTimeout = std::chrono::milliseconds{10000};
+
+    /** @brief Request timeout specified in milliseconds */
     std::chrono::milliseconds requestTimeout = std::chrono::milliseconds{0};  // no timeout at all
+
+    /** @brief Connection information; either ContactPoints or SecureConnectionBundle */
     std::variant<ContactPoints, SecureConnectionBundle> connectionInfo = ContactPoints{};
+
+    /** @brief The number of threads for the driver to pool */
     uint32_t threads = std::thread::hardware_concurrency();
-    uint32_t maxWriteRequestsOutstanding = 10'000;
-    uint32_t maxReadRequestsOutstanding = 100'000;
+
+    /** @brief The maximum number of outstanding write requests at any given moment */
+    uint32_t maxWriteRequestsOutstanding = 10'000u;
+
+    /** @brief The maximum number of outstanding read requests at any given moment */
+    uint32_t maxReadRequestsOutstanding = 100'000u;
+
+    /** @brief The number of connection per host to always have active */
+    uint32_t coreConnectionsPerHost = 1u;
+
+    /** @brief Size of the IO queue */
+    std::optional<uint32_t> queueSizeIO;
+
+    /** @brief SSL certificate */
     std::optional<std::string> certificate;  // ssl context
+
+    /** @brief Username/login */
     std::optional<std::string> username;
+
+    /** @brief Password to match the `username` */
     std::optional<std::string> password;
 
+    /**
+     * @brief Creates a new Settings object as a copy of the current one with overridden contact points.
+     */
     Settings
     withContactPoints(std::string_view contactPoints)
     {
@@ -65,6 +104,9 @@ struct Settings
         return tmp;
     }
 
+    /**
+     * @brief Returns the default settings.
+     */
     static Settings
     defaultSettings()
     {
@@ -74,7 +116,7 @@ struct Settings
 
 class Cluster : public ManagedObject<CassCluster>
 {
-    clio::Logger log_{"Backend"};
+    util::Logger log_{"Backend"};
 
 public:
     Cluster(Settings const& settings);
@@ -96,4 +138,4 @@ private:
     setupCredentials(Settings const& settings);
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/ExecutionStrategy.h

@@ -19,13 +19,13 @@
 
 #pragma once
 
-#include <backend/cassandra/Handle.h>
-#include <backend/cassandra/Types.h>
-#include <backend/cassandra/impl/AsyncExecutor.h>
-#include <log/Logger.h>
+#include <data/cassandra/Handle.h>
+#include <data/cassandra/Types.h>
+#include <data/cassandra/impl/AsyncExecutor.h>
 #include <util/Expected.h>
+#include <util/log/Logger.h>
 
-#include <boost/asio/async_result.hpp>
+#include <boost/asio.hpp>
 #include <boost/asio/spawn.hpp>
 
 #include <atomic>
@@ -36,19 +36,20 @@
 #include <optional>
 #include <thread>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
+
+// TODO: this could probably be also moved out of detail and into the main cassandra namespace.
 
 /**
- * @brief Implements async and sync querying against the cassandra DB with
- * support for throttling.
+ * @brief Implements async and sync querying against the cassandra DB with support for throttling.
  *
- * Note: A lot of the code that uses yield is repeated below. This is ok for now
- * because we are hopefully going to be getting rid of it entirely later on.
+ * Note: A lot of the code that uses yield is repeated below.
+ * This is ok for now because we are hopefully going to be getting rid of it entirely later on.
  */
 template <typename HandleType = Handle>
 class DefaultExecutionStrategy
 {
-    clio::Logger log_{"Backend"};
+    util::Logger log_{"Backend"};
 
     std::uint32_t maxWriteRequestsOutstanding_;
     std::atomic_uint32_t numWriteRequestsOutstanding_ = 0;
@@ -75,21 +76,21 @@ public:
     using FutureType = typename HandleType::FutureType;
     using FutureWithCallbackType = typename HandleType::FutureWithCallbackType;
     using ResultType = typename HandleType::ResultType;
-
     using CompletionTokenType = boost::asio::yield_context;
-    using FunctionType = void(boost::system::error_code);
-    using AsyncResultType = boost::asio::async_result<CompletionTokenType, FunctionType>;
-    using HandlerType = typename AsyncResultType::completion_handler_type;
 
-    DefaultExecutionStrategy(Settings settings, HandleType const& handle)
+    /**
+     * @param settings The settings to use
+     * @param handle A handle to the cassandra database
+     */
+    DefaultExecutionStrategy(Settings const& settings, HandleType const& handle)
         : maxWriteRequestsOutstanding_{settings.maxWriteRequestsOutstanding}
         , maxReadRequestsOutstanding_{settings.maxReadRequestsOutstanding}
         , work_{ioc_}
         , handle_{std::cref(handle)}
         , thread_{[this]() { ioc_.run(); }}
     {
-        log_.info() << "Max write requests outstanding is " << maxWriteRequestsOutstanding_
-                    << "; Max read requests outstanding is " << maxReadRequestsOutstanding_;
+        LOG(log_.info()) << "Max write requests outstanding is " << maxWriteRequestsOutstanding_
+                         << "; Max read requests outstanding is " << maxReadRequestsOutstanding_;
     }
 
     ~DefaultExecutionStrategy()
@@ -100,17 +101,20 @@ public:
     }
 
     /**
-     * @brief Wait for all async writes to finish before unblocking
+     * @brief Wait for all async writes to finish before unblocking.
      */
     void
     sync()
     {
-        log_.debug() << "Waiting to sync all writes...";
+        LOG(log_.debug()) << "Waiting to sync all writes...";
         std::unique_lock<std::mutex> lck(syncMutex_);
         syncCv_.wait(lck, [this]() { return finishedAllWriteRequests(); });
-        log_.debug() << "Sync done.";
+        LOG(log_.debug()) << "Sync done.";
     }
 
+    /**
+     * @return true if outstanding read requests allowance is exhausted; false otherwise
+     */
     bool
     isTooBusy() const
     {
@@ -118,7 +122,7 @@ public:
     }
 
     /**
-     * @brief Blocking query execution used for writing data
+     * @brief Blocking query execution used for writing data.
      *
      * Retries forever sleeping for 5 milliseconds between attempts.
      */
@@ -133,14 +137,14 @@ public:
             }
             else
             {
-                log_.warn() << "Cassandra sync write error, retrying: " << res.error();
+                LOG(log_.warn()) << "Cassandra sync write error, retrying: " << res.error();
                 std::this_thread::sleep_for(std::chrono::milliseconds(5));
             }
         }
     }
 
     /**
-     * @brief Blocking query execution used for writing data
+     * @brief Blocking query execution used for writing data.
      *
      * Retries forever sleeping for 5 milliseconds between attempts.
      */
@@ -152,11 +156,11 @@ public:
     }
 
     /**
-     * @brief Non-blocking query execution used for writing data
+     * @brief Non-blocking query execution used for writing data.
      *
      * Retries forever with retry policy specified by @ref AsyncExecutor
      *
-     * @param prepradeStatement Statement to prepare and execute
+     * @param preparedStatement Statement to prepare and execute
      * @param args Args to bind to the prepared statement
      * @throw DatabaseTimeout on timeout
      */
@@ -173,7 +177,7 @@ public:
     }
 
     /**
-     * @brief Non-blocking batched query execution used for writing data
+     * @brief Non-blocking batched query execution used for writing data.
      *
      * Retries forever with retry policy specified by @ref AsyncExecutor.
      *
@@ -199,7 +203,7 @@ public:
      * Retries forever until successful or throws an exception on timeout.
      *
      * @param token Completion token (yield_context)
-     * @param prepradeStatement Statement to prepare and execute
+     * @param preparedStatement Statement to prepare and execute
      * @param args Args to bind to the prepared statement
      * @throw DatabaseTimeout on timeout
      * @return ResultType or error wrapped in Expected
@@ -224,35 +228,35 @@ public:
     [[maybe_unused]] ResultOrErrorType
     read(CompletionTokenType token, std::vector<StatementType> const& statements)
     {
-        auto handler = HandlerType{token};
-        auto result = AsyncResultType{handler};
         auto const numStatements = statements.size();
+        std::optional<FutureWithCallbackType> future;
 
         // todo: perhaps use policy instead
         while (true)
         {
             numReadRequestsOutstanding_ += numStatements;
 
-            auto const future = handle_.get().asyncExecute(statements, [handler](auto&&) mutable {
-                boost::asio::post(boost::asio::get_associated_executor(handler), [handler]() mutable {
-                    handler(boost::system::error_code{});
-                });
-            });
+            auto init = [this, &statements, &future]<typename Self>(Self& self) {
+                auto sself = std::make_shared<Self>(std::move(self));
 
-            // suspend coroutine until completion handler is called
-            result.get();
+                future.emplace(handle_.get().asyncExecute(statements, [sself](auto&& res) mutable {
+                    boost::asio::post(
+                        boost::asio::get_associated_executor(*sself),
+                        [sself, res = std::move(res)]() mutable { sself->complete(std::move(res)); });
+                }));
+            };
 
+            auto res = boost::asio::async_compose<CompletionTokenType, void(ResultOrErrorType)>(
+                init, token, boost::asio::get_associated_executor(token));
             numReadRequestsOutstanding_ -= numStatements;
 
-            // it's safe to call blocking get on future here as we already
-            // waited for the coroutine to resume above.
-            if (auto res = future.get(); res)
+            if (res)
             {
                 return res;
             }
             else
             {
-                log_.error() << "Failed batch read in coroutine: " << res.error();
+                LOG(log_.error()) << "Failed batch read in coroutine: " << res.error();
                 throwErrorIfNeeded(res.error());
             }
         }
@@ -271,34 +275,33 @@ public:
     [[maybe_unused]] ResultOrErrorType
     read(CompletionTokenType token, StatementType const& statement)
     {
-        auto handler = HandlerType{token};
-        auto result = AsyncResultType{handler};
+        std::optional<FutureWithCallbackType> future;
 
         // todo: perhaps use policy instead
         while (true)
         {
             ++numReadRequestsOutstanding_;
+            auto init = [this, &statement, &future]<typename Self>(Self& self) {
+                auto sself = std::make_shared<Self>(std::move(self));
 
-            auto const future = handle_.get().asyncExecute(statement, [handler](auto const&) mutable {
-                boost::asio::post(boost::asio::get_associated_executor(handler), [handler]() mutable {
-                    handler(boost::system::error_code{});
-                });
-            });
-
-            // suspend coroutine until completion handler is called
-            result.get();
+                future.emplace(handle_.get().asyncExecute(statement, [sself](auto&& res) mutable {
+                    boost::asio::post(
+                        boost::asio::get_associated_executor(*sself),
+                        [sself, res = std::move(res)]() mutable { sself->complete(std::move(res)); });
+                }));
+            };
 
+            auto res = boost::asio::async_compose<CompletionTokenType, void(ResultOrErrorType)>(
+                init, token, boost::asio::get_associated_executor(token));
             --numReadRequestsOutstanding_;
 
-            // it's safe to call blocking get on future here as we already
-            // waited for the coroutine to resume above.
-            if (auto res = future.get(); res)
+            if (res)
             {
                 return res;
             }
             else
             {
-                log_.error() << "Failed read in coroutine: " << res.error();
+                LOG(log_.error()) << "Failed read in coroutine: " << res.error();
                 throwErrorIfNeeded(res.error());
             }
         }
@@ -318,9 +321,6 @@ public:
     std::vector<ResultType>
     readEach(CompletionTokenType token, std::vector<StatementType> const& statements)
     {
-        auto handler = HandlerType{token};
-        auto result = AsyncResultType{handler};
-
         std::atomic_bool hadError = false;
         std::atomic_int numOutstanding = statements.size();
         numReadRequestsOutstanding_ += statements.size();
@@ -328,29 +328,29 @@ public:
         auto futures = std::vector<FutureWithCallbackType>{};
         futures.reserve(numOutstanding);
 
-        // used as the handler for each async statement individually
-        auto executionHandler = [handler, &hadError, &numOutstanding](auto const& res) mutable {
-            if (not res)
-                hadError = true;
+        auto init = [this, &statements, &futures, &hadError, &numOutstanding]<typename Self>(Self& self) {
+            auto sself = std::make_shared<Self>(std::move(self));
+            auto executionHandler = [&hadError, &numOutstanding, sself](auto const& res) mutable {
+                if (not res)
+                    hadError = true;
 
-            // when all async operations complete unblock the result
-            if (--numOutstanding == 0)
-                boost::asio::post(boost::asio::get_associated_executor(handler), [handler]() mutable {
-                    handler(boost::system::error_code{});
+                // when all async operations complete unblock the result
+                if (--numOutstanding == 0)
+                    boost::asio::post(
+                        boost::asio::get_associated_executor(*sself), [sself]() mutable { sself->complete(); });
+            };
+
+            std::transform(
+                std::cbegin(statements),
+                std::cend(statements),
+                std::back_inserter(futures),
+                [this, &executionHandler](auto const& statement) {
+                    return handle_.get().asyncExecute(statement, executionHandler);
                 });
         };
 
-        std::transform(
-            std::cbegin(statements),
-            std::cend(statements),
-            std::back_inserter(futures),
-            [this, &executionHandler](auto const& statement) {
-                return handle_.get().asyncExecute(statement, executionHandler);
-            });
-
-        // suspend coroutine until completion handler is called
-        result.get();
-
+        boost::asio::async_compose<CompletionTokenType, void()>(
+            init, token, boost::asio::get_associated_executor(token));
         numReadRequestsOutstanding_ -= statements.size();
 
         if (hadError)
@@ -359,8 +359,7 @@ public:
         std::vector<ResultType> results;
         results.reserve(futures.size());
 
-        // it's safe to call blocking get on futures here as we already
-        // waited for the coroutine to resume above.
+        // it's safe to call blocking get on futures here as we already waited for the coroutine to resume above.
         std::transform(
             std::make_move_iterator(std::begin(futures)),
             std::make_move_iterator(std::end(futures)),
@@ -384,8 +383,8 @@ private:
             std::unique_lock<std::mutex> lck(throttleMutex_);
             if (!canAddWriteRequest())
             {
-                log_.trace() << "Max outstanding requests reached. "
-                             << "Waiting for other requests to finish";
+                LOG(log_.trace()) << "Max outstanding requests reached. "
+                                  << "Waiting for other requests to finish";
                 throttleCv_.wait(lck, [this]() { return canAddWriteRequest(); });
             }
         }
@@ -440,4 +439,4 @@ private:
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Future.cpp

@@ -17,9 +17,9 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/Error.h>
-#include <backend/cassandra/impl/Future.h>
-#include <backend/cassandra/impl/Result.h>
+#include <data/cassandra/Error.h>
+#include <data/cassandra/impl/Future.h>
+#include <data/cassandra/impl/Result.h>
 
 #include <exception>
 #include <vector>
@@ -28,7 +28,7 @@ namespace {
 static constexpr auto futureDeleter = [](CassFuture* ptr) { cass_future_free(ptr); };
 }  // namespace
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 /* implicit */ Future::Future(CassFuture* ptr) : ManagedObject{ptr, futureDeleter}
 {
@@ -73,7 +73,10 @@ void
 invokeHelper(CassFuture* ptr, void* cbPtr)
 {
     // Note: can't use Future{ptr}.get() because double free will occur :/
-    auto* cb = static_cast<FutureWithCallback::fn_t*>(cbPtr);
+    // Note2: we are moving/copying it locally as a workaround for an issue we are seeing from asio recently.
+    // stackoverflow.com/questions/77004137/boost-asio-async-compose-gets-stuck-under-load
+    auto* cb = static_cast<FutureWithCallback::FnType*>(cbPtr);
+    auto local = std::make_unique<FutureWithCallback::FnType>(std::move(*cb));
     if (auto const rc = cass_future_error_code(ptr); rc)
     {
         auto const errMsg = [&ptr](std::string const& label) {
@@ -82,16 +85,16 @@ invokeHelper(CassFuture* ptr, void* cbPtr)
             cass_future_error_message(ptr, &message, &len);
             return label + ": " + std::string{message, len};
         }("invokeHelper");
-        (*cb)(Error{CassandraError{errMsg, rc}});
+        (*local)(Error{CassandraError{errMsg, rc}});
     }
     else
     {
-        (*cb)(Result{cass_future_get_result(ptr)});
+        (*local)(Result{cass_future_get_result(ptr)});
     }
 }
 
-/* implicit */ FutureWithCallback::FutureWithCallback(CassFuture* ptr, fn_t&& cb)
-    : Future{ptr}, cb_{std::make_unique<fn_t>(std::move(cb))}
+/* implicit */ FutureWithCallback::FutureWithCallback(CassFuture* ptr, FnType&& cb)
+    : Future{ptr}, cb_{std::make_unique<FnType>(std::move(cb))}
 {
     // Instead of passing `this` as the userdata void*, we pass the address of
     // the callback itself which will survive std::move of the
@@ -99,4 +102,4 @@ invokeHelper(CassFuture* ptr, void* cbPtr)
     cass_future_set_callback(*this, &invokeHelper, cb_.get());
 }
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Future.h

@@ -19,12 +19,12 @@
 
 #pragma once
 
-#include <backend/cassandra/Types.h>
-#include <backend/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/Types.h>
+#include <data/cassandra/impl/ManagedObject.h>
 
 #include <cassandra.h>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 struct Future : public ManagedObject<CassFuture>
 {
@@ -43,16 +43,16 @@ invokeHelper(CassFuture* ptr, void* self);
 class FutureWithCallback : public Future
 {
 public:
-    using fn_t = std::function<void(ResultOrError)>;
-    using fn_ptr_t = std::unique_ptr<fn_t>;
+    using FnType = std::function<void(ResultOrError)>;
+    using FnPtrType = std::unique_ptr<FnType>;
 
-    /* implicit */ FutureWithCallback(CassFuture* ptr, fn_t&& cb);
+    /* implicit */ FutureWithCallback(CassFuture* ptr, FnType&& cb);
     FutureWithCallback(FutureWithCallback const&) = delete;
     FutureWithCallback(FutureWithCallback&&) = default;
 
 private:
-    /*! Wrapped in a unique_ptr so it can survive std::move :/ */
-    fn_ptr_t cb_;
+    /** Wrapped in a unique_ptr so it can survive std::move :/ */
+    FnPtrType cb_;
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/ManagedObject.h

@@ -21,7 +21,7 @@
 
 #include <memory>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 template <typename Managed>
 class ManagedObject
@@ -38,10 +38,10 @@ public:
     }
     ManagedObject(ManagedObject&&) = default;
 
-    operator Managed* const() const
+    operator Managed*() const
     {
         return ptr_.get();
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Result.cpp

@@ -17,14 +17,14 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/impl/Result.h>
+#include <data/cassandra/impl/Result.h>
 
 namespace {
 static constexpr auto resultDeleter = [](CassResult const* ptr) { cass_result_free(ptr); };
 static constexpr auto resultIteratorDeleter = [](CassIterator* ptr) { cass_iterator_free(ptr); };
 }  // namespace
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 /* implicit */ Result::Result(CassResult const* ptr) : ManagedObject{ptr, resultDeleter}
 {
@@ -66,4 +66,4 @@ ResultIterator::hasMore() const
     return hasMore_;
 }
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Result.h

@@ -19,8 +19,8 @@
 
 #pragma once
 
-#include <backend/cassandra/impl/ManagedObject.h>
-#include <backend/cassandra/impl/Tuple.h>
+#include <data/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/impl/Tuple.h>
 #include <util/Expected.h>
 
 #include <ripple/basics/base_uint.h>
@@ -31,7 +31,7 @@
 #include <iterator>
 #include <tuple>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 template <typename>
 static constexpr bool unsupported_v = false;
@@ -51,11 +51,11 @@ extractColumn(CassRow const* row, std::size_t idx)
         }
     };
 
-    using decayed_t = std::decay_t<Type>;
-    using uint_tuple_t = std::tuple<uint32_t, uint32_t>;
-    using uchar_vector_t = std::vector<unsigned char>;
+    using DecayedType = std::decay_t<Type>;
+    using UintTupleType = std::tuple<uint32_t, uint32_t>;
+    using UCharVectorType = std::vector<unsigned char>;
 
-    if constexpr (std::is_same_v<decayed_t, ripple::uint256>)
+    if constexpr (std::is_same_v<DecayedType, ripple::uint256>)
     {
         cass_byte_t const* buf;
         std::size_t bufSize;
@@ -63,7 +63,7 @@ extractColumn(CassRow const* row, std::size_t idx)
         throwErrorIfNeeded(rc, "Extract ripple::uint256");
         output = ripple::uint256::fromVoid(buf);
     }
-    else if constexpr (std::is_same_v<decayed_t, ripple::AccountID>)
+    else if constexpr (std::is_same_v<DecayedType, ripple::AccountID>)
     {
         cass_byte_t const* buf;
         std::size_t bufSize;
@@ -71,20 +71,20 @@ extractColumn(CassRow const* row, std::size_t idx)
         throwErrorIfNeeded(rc, "Extract ripple::AccountID");
         output = ripple::AccountID::fromVoid(buf);
     }
-    else if constexpr (std::is_same_v<decayed_t, uchar_vector_t>)
+    else if constexpr (std::is_same_v<DecayedType, UCharVectorType>)
     {
         cass_byte_t const* buf;
         std::size_t bufSize;
         auto const rc = cass_value_get_bytes(cass_row_get_column(row, idx), &buf, &bufSize);
         throwErrorIfNeeded(rc, "Extract vector<unsigned char>");
-        output = uchar_vector_t{buf, buf + bufSize};
+        output = UCharVectorType{buf, buf + bufSize};
     }
-    else if constexpr (std::is_same_v<decayed_t, uint_tuple_t>)
+    else if constexpr (std::is_same_v<DecayedType, UintTupleType>)
     {
         auto const* tuple = cass_row_get_column(row, idx);
         output = TupleIterator::fromTuple(tuple).extract<uint32_t, uint32_t>();
     }
-    else if constexpr (std::is_convertible_v<decayed_t, std::string>)
+    else if constexpr (std::is_convertible_v<DecayedType, std::string>)
     {
         char const* value;
         std::size_t len;
@@ -92,7 +92,7 @@ extractColumn(CassRow const* row, std::size_t idx)
         throwErrorIfNeeded(rc, "Extract string");
         output = std::string{value, len};
     }
-    else if constexpr (std::is_same_v<decayed_t, bool>)
+    else if constexpr (std::is_same_v<DecayedType, bool>)
     {
         cass_bool_t flag;
         auto const rc = cass_value_get_bool(cass_row_get_column(row, idx), &flag);
@@ -100,17 +100,17 @@ extractColumn(CassRow const* row, std::size_t idx)
         output = flag ? true : false;
     }
     // clio only uses bigint (int64_t) so we convert any incoming type
-    else if constexpr (std::is_convertible_v<decayed_t, int64_t>)
+    else if constexpr (std::is_convertible_v<DecayedType, int64_t>)
     {
         int64_t out;
         auto const rc = cass_value_get_int64(cass_row_get_column(row, idx), &out);
         throwErrorIfNeeded(rc, "Extract int64");
-        output = static_cast<decayed_t>(out);
+        output = static_cast<DecayedType>(out);
     }
     else
     {
         // type not supported for extraction
-        static_assert(unsupported_v<decayed_t>);
+        static_assert(unsupported_v<DecayedType>);
     }
 
     return output;
@@ -254,4 +254,4 @@ public:
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/RetryPolicy.h

@@ -19,10 +19,10 @@
 
 #pragma once
 
-#include <backend/cassandra/Handle.h>
-#include <backend/cassandra/Types.h>
-#include <log/Logger.h>
+#include <data/cassandra/Handle.h>
+#include <data/cassandra/Types.h>
 #include <util/Expected.h>
+#include <util/log/Logger.h>
 
 #include <boost/asio.hpp>
 
@@ -30,14 +30,14 @@
 #include <chrono>
 #include <cmath>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 /**
  * @brief A retry policy that employs exponential backoff
  */
 class ExponentialBackoffRetryPolicy
 {
-    clio::Logger log_{"Backend"};
+    util::Logger log_{"Backend"};
 
     boost::asio::steady_timer timer_;
     uint32_t attempt_ = 0u;
@@ -46,7 +46,7 @@ public:
     /**
      * @brief Create a new retry policy instance with the io_context provided
      */
-    ExponentialBackoffRetryPolicy(boost::asio::io_context& ioc) : timer_{ioc}
+    ExponentialBackoffRetryPolicy(boost::asio::io_context& ioc) : timer_{boost::asio::make_strand(ioc)}
     {
     }
 
@@ -59,8 +59,8 @@ public:
     shouldRetry([[maybe_unused]] CassandraError err)
     {
         auto const delay = calculateDelay(attempt_);
-        log_.error() << "Cassandra write error: " << err << ", current retries " << attempt_ << ", retrying in "
-                     << delay.count() << " milliseconds";
+        LOG(log_.error()) << "Cassandra write error: " << err << ", current retries " << attempt_ << ", retrying in "
+                          << delay.count() << " milliseconds";
 
         return true;  // keep retrying forever
     }
@@ -91,4 +91,4 @@ public:
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Session.h

@@ -19,11 +19,11 @@
 
 #pragma once
 
-#include <backend/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/impl/ManagedObject.h>
 
 #include <cassandra.h>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 class Session : public ManagedObject<CassSession>
 {
@@ -35,4 +35,4 @@ public:
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/SslContext.cpp

@@ -17,13 +17,13 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/impl/SslContext.h>
+#include <data/cassandra/impl/SslContext.h>
 
 namespace {
 static constexpr auto contextDeleter = [](CassSsl* ptr) { cass_ssl_free(ptr); };
 }  // namespace
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 SslContext::SslContext(std::string const& certificate) : ManagedObject{cass_ssl_new(), contextDeleter}
 {
@@ -34,4 +34,4 @@ SslContext::SslContext(std::string const& certificate) : ManagedObject{cass_ssl_
     }
 }
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/SslContext.h

@@ -19,17 +19,17 @@
 
 #pragma once
 
-#include <backend/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/impl/ManagedObject.h>
 
 #include <cassandra.h>
 
 #include <string>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 struct SslContext : public ManagedObject<CassSsl>
 {
     explicit SslContext(std::string const& certificate);
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Statement.h

@@ -19,9 +19,9 @@
 
 #pragma once
 
-#include <backend/cassandra/Types.h>
-#include <backend/cassandra/impl/ManagedObject.h>
-#include <backend/cassandra/impl/Tuple.h>
+#include <data/cassandra/Types.h>
+#include <data/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/impl/Tuple.h>
 #include <util/Expected.h>
 
 #include <ripple/basics/base_uint.h>
@@ -33,7 +33,7 @@
 #include <compare>
 #include <iterator>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 class Statement : public ManagedObject<CassStatement>
 {
@@ -44,7 +44,7 @@ class Statement : public ManagedObject<CassStatement>
 
 public:
     /**
-     * @brief Construct a new statement with optionally provided arguments
+     * @brief Construct a new statement with optionally provided arguments.
      *
      * Note: it's up to the user to make sure the bound parameters match
      * the format of the query (e.g. amount of '?' matches count of args).
@@ -66,6 +66,11 @@ public:
 
     Statement(Statement&&) = default;
 
+    /**
+     * @brief Binds the given arguments to the statement.
+     *
+     * @param args Arguments to bind
+     */
     template <typename... Args>
     void
     bind(Args&&... args) const
@@ -74,6 +79,12 @@ public:
         (this->bindAt<Args>(idx++, std::forward<Args>(args)), ...);
     }
 
+    /**
+     * @brief Binds an argument to a specific index.
+     *
+     * @param idx The index of the argument
+     * @param value The value to bind it to
+     */
     template <typename Type>
     void
     bindAt(std::size_t const idx, Type&& value) const
@@ -88,48 +99,48 @@ public:
             return cass_statement_bind_bytes(*this, idx, static_cast<cass_byte_t const*>(data), size);
         };
 
-        using decayed_t = std::decay_t<Type>;
-        using uchar_vec_t = std::vector<unsigned char>;
-        using uint_tuple_t = std::tuple<uint32_t, uint32_t>;
+        using DecayedType = std::decay_t<Type>;
+        using UCharVectorType = std::vector<unsigned char>;
+        using UintTupleType = std::tuple<uint32_t, uint32_t>;
 
-        if constexpr (std::is_same_v<decayed_t, ripple::uint256>)
+        if constexpr (std::is_same_v<DecayedType, ripple::uint256>)
         {
             auto const rc = bindBytes(value.data(), value.size());
             throwErrorIfNeeded(rc, "Bind ripple::uint256");
         }
-        else if constexpr (std::is_same_v<decayed_t, ripple::AccountID>)
+        else if constexpr (std::is_same_v<DecayedType, ripple::AccountID>)
         {
             auto const rc = bindBytes(value.data(), value.size());
             throwErrorIfNeeded(rc, "Bind ripple::AccountID");
         }
-        else if constexpr (std::is_same_v<decayed_t, uchar_vec_t>)
+        else if constexpr (std::is_same_v<DecayedType, UCharVectorType>)
         {
             auto const rc = bindBytes(value.data(), value.size());
             throwErrorIfNeeded(rc, "Bind vector<unsigned char>");
         }
-        else if constexpr (std::is_convertible_v<decayed_t, std::string>)
+        else if constexpr (std::is_convertible_v<DecayedType, std::string>)
         {
             // reinterpret_cast is needed here :'(
             auto const rc = bindBytes(reinterpret_cast<unsigned char const*>(value.data()), value.size());
             throwErrorIfNeeded(rc, "Bind string (as bytes)");
         }
-        else if constexpr (std::is_same_v<decayed_t, uint_tuple_t>)
+        else if constexpr (std::is_same_v<DecayedType, UintTupleType>)
         {
             auto const rc = cass_statement_bind_tuple(*this, idx, Tuple{std::move(value)});
             throwErrorIfNeeded(rc, "Bind tuple<uint32, uint32>");
         }
-        else if constexpr (std::is_same_v<decayed_t, bool>)
+        else if constexpr (std::is_same_v<DecayedType, bool>)
         {
             auto const rc = cass_statement_bind_bool(*this, idx, value ? cass_true : cass_false);
             throwErrorIfNeeded(rc, "Bind bool");
         }
-        else if constexpr (std::is_same_v<decayed_t, Limit>)
+        else if constexpr (std::is_same_v<DecayedType, Limit>)
         {
             auto const rc = cass_statement_bind_int32(*this, idx, value.limit);
             throwErrorIfNeeded(rc, "Bind limit (int32)");
         }
         // clio only uses bigint (int64_t) so we convert any incoming type
-        else if constexpr (std::is_convertible_v<decayed_t, int64_t>)
+        else if constexpr (std::is_convertible_v<DecayedType, int64_t>)
         {
             auto const rc = cass_statement_bind_int64(*this, idx, value);
             throwErrorIfNeeded(rc, "Bind int64");
@@ -137,11 +148,16 @@ public:
         else
         {
             // type not supported for binding
-            static_assert(unsupported_v<decayed_t>);
+            static_assert(unsupported_v<DecayedType>);
         }
     }
 };
 
+/**
+ * @brief Represents a prepared statement on the DB side.
+ *
+ * This is used to produce Statement objects that can be executed.
+ */
 class PreparedStatement : public ManagedObject<CassPrepared const>
 {
     static constexpr auto deleter = [](CassPrepared const* ptr) { cass_prepared_free(ptr); };
@@ -151,6 +167,12 @@ public:
     {
     }
 
+    /**
+     * @brief Bind the given arguments and produce a ready to execute Statement.
+     *
+     * @param args The arguments to bind
+     * @return A bound and ready to execute Statement object
+     */
     template <typename... Args>
     Statement
     bind(Args&&... args) const
@@ -161,4 +183,4 @@ public:
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Tuple.cpp

@@ -17,14 +17,14 @@
 */
 //==============================================================================
 
-#include <backend/cassandra/impl/Tuple.h>
+#include <data/cassandra/impl/Tuple.h>
 
 namespace {
 static constexpr auto tupleDeleter = [](CassTuple* ptr) { cass_tuple_free(ptr); };
 static constexpr auto tupleIteratorDeleter = [](CassIterator* ptr) { cass_iterator_free(ptr); };
 }  // namespace
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 /* implicit */ Tuple::Tuple(CassTuple* ptr) : ManagedObject{ptr, tupleDeleter}
 {
@@ -40,4 +40,4 @@ TupleIterator::fromTuple(CassValue const* value)
     return {cass_iterator_from_tuple(value)};
 }
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/data/cassandra/impl/Tuple.h

@@ -19,7 +19,7 @@
 
 #pragma once
 
-#include <backend/cassandra/impl/ManagedObject.h>
+#include <data/cassandra/impl/ManagedObject.h>
 
 #include <cassandra.h>
 
@@ -28,7 +28,7 @@
 #include <string_view>
 #include <tuple>
 
-namespace Backend::Cassandra::detail {
+namespace data::cassandra::detail {
 
 class Tuple : public ManagedObject<CassTuple>
 {
@@ -68,15 +68,15 @@ public:
             }
         };
 
-        using decayed_t = std::decay_t<Type>;
+        using DecayedType = std::decay_t<Type>;
 
-        if constexpr (std::is_same_v<decayed_t, bool>)
+        if constexpr (std::is_same_v<DecayedType, bool>)
         {
             auto const rc = cass_tuple_set_bool(*this, idx, value ? cass_true : cass_false);
             throwErrorIfNeeded(rc, "Bind bool");
         }
         // clio only uses bigint (int64_t) so we convert any incoming type
-        else if constexpr (std::is_convertible_v<decayed_t, int64_t>)
+        else if constexpr (std::is_convertible_v<DecayedType, int64_t>)
         {
             auto const rc = cass_tuple_set_int64(*this, idx, value);
             throwErrorIfNeeded(rc, "Bind int64");
@@ -84,7 +84,7 @@ public:
         else
         {
             // type not supported for binding
-            static_assert(unsupported_v<decayed_t>);
+            static_assert(unsupported_v<DecayedType>);
         }
     }
 };
@@ -126,24 +126,24 @@ private:
             }
         };
 
-        using decayed_t = std::decay_t<Type>;
+        using DecayedType = std::decay_t<Type>;
 
         // clio only uses bigint (int64_t) so we convert any incoming type
-        if constexpr (std::is_convertible_v<decayed_t, int64_t>)
+        if constexpr (std::is_convertible_v<DecayedType, int64_t>)
         {
             int64_t out;
             auto const rc = cass_value_get_int64(cass_iterator_get_value(*this), &out);
             throwErrorIfNeeded(rc, "Extract int64 from tuple");
-            output = static_cast<decayed_t>(out);
+            output = static_cast<DecayedType>(out);
         }
         else
         {
             // type not supported for extraction
-            static_assert(unsupported_v<decayed_t>);
+            static_assert(unsupported_v<DecayedType>);
         }
 
         return output;
     }
 };
 
-}  // namespace Backend::Cassandra::detail
+}  // namespace data::cassandra::detail

src/etl/ETLHelpers.h

@@ -17,6 +17,7 @@
 */
 //==============================================================================
 
+/** @file */
 #pragma once
 
 #include <ripple/basics/base_uint.h>
@@ -26,6 +27,7 @@
 #include <queue>
 #include <sstream>
 
+namespace etl {
 /**
  * @brief This datastructure is used to keep track of the sequence of the most recent ledger validated by the network.
  *
@@ -43,6 +45,9 @@ class NetworkValidatedLedgers
     std::condition_variable cv_;
 
 public:
+    /**
+     * @brief A factory function for NetworkValidatedLedgers.
+     */
     static std::shared_ptr<NetworkValidatedLedgers>
     make_ValidatedLedgers()
     {
@@ -50,9 +55,9 @@ public:
     }
 
     /**
-     * @brief Notify the datastructure that idx has been validated by the network
+     * @brief Notify the datastructure that idx has been validated by the network.
      *
-     * @param idx sequence validated by network
+     * @param idx Sequence validated by network
      */
     void
     push(uint32_t idx)
@@ -68,7 +73,7 @@ public:
      *
      * If no ledgers are known to have been validated, this function waits until the next ledger is validated
      *
-     * @return sequence of most recently validated ledger. empty optional if the datastructure has been stopped
+     * @return Sequence of most recently validated ledger. empty optional if the datastructure has been stopped
      */
     std::optional<uint32_t>
     getMostRecent()
@@ -79,9 +84,9 @@ public:
     }
 
     /**
-     * @brief Waits for the sequence to be validated by the network
+     * @brief Waits for the sequence to be validated by the network.
      *
-     * @param sequence to wait for
+     * @param sequence The sequence to wait for
      * @return true if sequence was validated, false otherwise a return value of false means the datastructure has been
      * stopped
      */
@@ -100,7 +105,7 @@ public:
 
 // TODO: does the note make sense? lockfree queues provide the same blocking behaviour just without mutex, don't they?
 /**
- * @brief Generic thread-safe queue with a max capacity
+ * @brief Generic thread-safe queue with a max capacity.
  *
  * @note (original note) We can't use a lockfree queue here, since we need the ability to wait for an element to be
  * added or removed from the queue. These waits are blocking calls.
@@ -116,21 +121,21 @@ class ThreadSafeQueue
 
 public:
     /**
-     * @brief Create an instance of the queue
+     * @brief Create an instance of the queue.
      *
      * @param maxSize maximum size of the queue. Calls that would cause the queue to exceed this size will block until
-     * free space is available
+     * free space is available.
      */
     ThreadSafeQueue(uint32_t maxSize) : maxSize_(maxSize)
     {
     }
 
     /**
-     * @brief Push element onto the queue
+     * @brief Push element onto the queue.
      *
-     * Note: This method will block until free space is available
+     * Note: This method will block until free space is available.
      *
-     * @param elt element to push onto queue
+     * @param elt Element to push onto queue
      */
     void
     push(T const& elt)
@@ -142,11 +147,11 @@ public:
     }
 
     /**
-     * @brief Push element onto the queue
+     * @brief Push element onto the queue.
      *
      * Note: This method will block until free space is available
      *
-     * @param elt element to push onto queue. elt is moved from
+     * @param elt Element to push onto queue. Ownership is transferred
      */
     void
     push(T&& elt)
@@ -158,11 +163,11 @@ public:
     }
 
     /**
-     * @brief Pop element from the queue
+     * @brief Pop element from the queue.
      *
-     * Note: Will block until queue is non-empty
+     * Note: Will block until queue is non-empty.
      *
-     * @return element popped from queue
+     * @return Element popped from queue
      */
     T
     pop()
@@ -178,9 +183,9 @@ public:
     }
 
     /**
-     * @brief Attempt to pop an element
+     * @brief Attempt to pop an element.
      *
-     * @return element popped from queue or empty optional if queue was empty
+     * @return Element popped from queue or empty optional if queue was empty
      */
     std::optional<T>
     tryPop()
@@ -200,7 +205,7 @@ public:
 /**
  * @brief Parititions the uint256 keyspace into numMarkers partitions, each of equal size.
  *
- * @param numMarkers total markers to partition for
+ * @param numMarkers Total markers to partition for
  */
 inline std::vector<ripple::uint256>
 getMarkers(size_t numMarkers)
@@ -219,3 +224,4 @@ getMarkers(size_t numMarkers)
     }
     return markers;
 }
+}  // namespace etl

src/etl/ETLService.cpp

@@ -19,16 +19,17 @@
 
 #include <etl/ETLService.h>
 
-using namespace clio;
+#include <ripple/protocol/LedgerHeader.h>
 
+namespace etl {
 // Database must be populated when this starts
 std::optional<uint32_t>
-ETLService::runETLPipeline(uint32_t startSequence, int numExtractors)
+ETLService::runETLPipeline(uint32_t startSequence, uint32_t numExtractors)
 {
     if (finishSequence_ && startSequence > *finishSequence_)
         return {};
 
-    log_.debug() << "Starting etl pipeline";
+    LOG(log_.debug()) << "Starting etl pipeline";
     state_.isWriting = true;
 
     auto rng = backend_->hardFetchLedgerRangeNoThrow();
@@ -46,7 +47,8 @@ ETLService::runETLPipeline(uint32_t startSequence, int numExtractors)
         extractors.push_back(std::make_unique<ExtractorType>(
             pipe, networkValidatedLedgers_, ledgerFetcher_, startSequence + i, finishSequence_, state_));
 
-    auto transformer = TransformerType{pipe, backend_, ledgerLoader_, ledgerPublisher_, startSequence, state_};
+    auto transformer =
+        TransformerType{pipe, backend_, ledgerLoader_, ledgerPublisher_, amendmentBlockHandler_, startSequence, state_};
     transformer.waitTillFinished();  // suspend current thread until exit condition is met
     pipe.cleanup();                  // TODO: this should probably happen automatically using destructor
 
@@ -56,12 +58,12 @@ ETLService::runETLPipeline(uint32_t startSequence, int numExtractors)
 
     auto const end = std::chrono::system_clock::now();
     auto const lastPublishedSeq = ledgerPublisher_.getLastPublishedSequence();
-    log_.debug() << "Extracted and wrote " << lastPublishedSeq.value_or(startSequence) - startSequence << " in "
-                 << ((end - begin).count()) / 1000000000.0;
+    LOG(log_.debug()) << "Extracted and wrote " << lastPublishedSeq.value_or(startSequence) - startSequence << " in "
+                      << ((end - begin).count()) / 1000000000.0;
 
     state_.isWriting = false;
 
-    log_.debug() << "Stopping etl pipeline";
+    LOG(log_.debug()) << "Stopping etl pipeline";
     return lastPublishedSeq;
 }
 
@@ -79,35 +81,39 @@ ETLService::monitor()
     auto rng = backend_->hardFetchLedgerRangeNoThrow();
     if (!rng)
     {
-        log_.info() << "Database is empty. Will download a ledger "
-                       "from the network.";
-        std::optional<ripple::LedgerInfo> ledger;
+        LOG(log_.info()) << "Database is empty. Will download a ledger from the network.";
+        std::optional<ripple::LedgerHeader> ledger;
 
-        if (startSequence_)
+        try
         {
-            log_.info() << "ledger sequence specified in config. "
-                        << "Will begin ETL process starting with ledger " << *startSequence_;
-            ledger = ledgerLoader_.loadInitialLedger(*startSequence_);
-        }
-        else
-        {
-            log_.info() << "Waiting for next ledger to be validated by network...";
-            std::optional<uint32_t> mostRecentValidated = networkValidatedLedgers_->getMostRecent();
-
-            if (mostRecentValidated)
+            if (startSequence_)
             {
-                log_.info() << "Ledger " << *mostRecentValidated << " has been validated. "
-                            << "Downloading...";
-                ledger = ledgerLoader_.loadInitialLedger(*mostRecentValidated);
+                LOG(log_.info()) << "ledger sequence specified in config. "
+                                 << "Will begin ETL process starting with ledger " << *startSequence_;
+                ledger = ledgerLoader_.loadInitialLedger(*startSequence_);
             }
             else
             {
-                log_.info() << "The wait for the next validated "
-                            << "ledger has been aborted. "
-                            << "Exiting monitor loop";
-                return;
+                LOG(log_.info()) << "Waiting for next ledger to be validated by network...";
+                std::optional<uint32_t> mostRecentValidated = networkValidatedLedgers_->getMostRecent();
+
+                if (mostRecentValidated)
+                {
+                    LOG(log_.info()) << "Ledger " << *mostRecentValidated << " has been validated. Downloading...";
+                    ledger = ledgerLoader_.loadInitialLedger(*mostRecentValidated);
+                }
+                else
+                {
+                    LOG(log_.info()) << "The wait for the next validated ledger has been aborted. Exiting monitor loop";
+                    return;
+                }
             }
         }
+        catch (std::runtime_error const& e)
+        {
+            LOG(log_.fatal()) << "Failed to load initial ledger: " << e.what();
+            return amendmentBlockHandler_.onAmendmentBlock();
+        }
 
         if (ledger)
         {
@@ -115,86 +121,101 @@ ETLService::monitor()
         }
         else
         {
-            log_.error() << "Failed to load initial ledger. Exiting monitor loop";
+            LOG(log_.error()) << "Failed to load initial ledger. Exiting monitor loop";
             return;
         }
     }
     else
     {
         if (startSequence_)
-            log_.warn() << "start sequence specified but db is already populated";
+            LOG(log_.warn()) << "start sequence specified but db is already populated";
 
-        log_.info() << "Database already populated. Picking up from the tip of history";
+        LOG(log_.info()) << "Database already populated. Picking up from the tip of history";
         cacheLoader_.load(rng->maxSequence);
     }
 
     assert(rng);
     uint32_t nextSequence = rng->maxSequence + 1;
 
-    log_.debug() << "Database is populated. "
-                 << "Starting monitor loop. sequence = " << nextSequence;
+    LOG(log_.debug()) << "Database is populated. "
+                      << "Starting monitor loop. sequence = " << nextSequence;
 
     while (true)
     {
-        if (auto rng = backend_->hardFetchLedgerRangeNoThrow(); rng && rng->maxSequence >= nextSequence)
+        nextSequence = publishNextSequence(nextSequence);
+    }
+}
+
+uint32_t
+ETLService::publishNextSequence(uint32_t nextSequence)
+{
+    if (auto rng = backend_->hardFetchLedgerRangeNoThrow(); rng && rng->maxSequence >= nextSequence)
+    {
+        ledgerPublisher_.publish(nextSequence, {});
+        ++nextSequence;
+    }
+    else if (networkValidatedLedgers_->waitUntilValidatedByNetwork(nextSequence, 1000))
+    {
+        LOG(log_.info()) << "Ledger with sequence = " << nextSequence << " has been validated by the network. "
+                         << "Attempting to find in database and publish";
+
+        // Attempt to take over responsibility of ETL writer after 10 failed
+        // attempts to publish the ledger. publishLedger() fails if the
+        // ledger that has been validated by the network is not found in the
+        // database after the specified number of attempts. publishLedger()
+        // waits one second between each attempt to read the ledger from the
+        // database
+        constexpr size_t timeoutSeconds = 10;
+        bool success = ledgerPublisher_.publish(nextSequence, timeoutSeconds);
+
+        if (!success)
+        {
+            LOG(log_.warn()) << "Failed to publish ledger with sequence = " << nextSequence << " . Beginning ETL";
+
+            // returns the most recent sequence published empty optional if no sequence was published
+            std::optional<uint32_t> lastPublished = runETLPipeline(nextSequence, extractorThreads_);
+            LOG(log_.info()) << "Aborting ETL. Falling back to publishing";
+
+            // if no ledger was published, don't increment nextSequence
+            if (lastPublished)
+                nextSequence = *lastPublished + 1;
+        }
+        else
         {
-            ledgerPublisher_.publish(nextSequence, {});
             ++nextSequence;
         }
-        else if (networkValidatedLedgers_->waitUntilValidatedByNetwork(nextSequence, 1000))
-        {
-            log_.info() << "Ledger with sequence = " << nextSequence << " has been validated by the network. "
-                        << "Attempting to find in database and publish";
-
-            // Attempt to take over responsibility of ETL writer after 10 failed
-            // attempts to publish the ledger. publishLedger() fails if the
-            // ledger that has been validated by the network is not found in the
-            // database after the specified number of attempts. publishLedger()
-            // waits one second between each attempt to read the ledger from the
-            // database
-            constexpr size_t timeoutSeconds = 10;
-            bool success = ledgerPublisher_.publish(nextSequence, timeoutSeconds);
-
-            if (!success)
-            {
-                log_.warn() << "Failed to publish ledger with sequence = " << nextSequence << " . Beginning ETL";
-
-                // returns the most recent sequence published empty optional if no sequence was published
-                std::optional<uint32_t> lastPublished = runETLPipeline(nextSequence, extractorThreads_);
-                log_.info() << "Aborting ETL. Falling back to publishing";
-
-                // if no ledger was published, don't increment nextSequence
-                if (lastPublished)
-                    nextSequence = *lastPublished + 1;
-            }
-            else
-            {
-                ++nextSequence;
-            }
-        }
     }
+    return nextSequence;
 }
 
 void
 ETLService::monitorReadOnly()
 {
-    log_.debug() << "Starting reporting in strict read only mode";
+    LOG(log_.debug()) << "Starting reporting in strict read only mode";
 
-    auto rng = backend_->hardFetchLedgerRangeNoThrow();
-    uint32_t latestSequence;
+    const auto latestSequenceOpt = [this]() -> std::optional<uint32_t> {
+        auto rng = backend_->hardFetchLedgerRangeNoThrow();
 
-    if (!rng)
-    {
-        if (auto net = networkValidatedLedgers_->getMostRecent())
-            latestSequence = *net;
+        if (!rng)
+        {
+            if (auto net = networkValidatedLedgers_->getMostRecent())
+                return *net;
+            else
+                return std::nullopt;
+        }
         else
-            return;
-    }
-    else
+        {
+            return rng->maxSequence;
+        }
+    }();
+
+    if (!latestSequenceOpt.has_value())
     {
-        latestSequence = rng->maxSequence;
+        return;
     }
 
+    uint32_t latestSequence = *latestSequenceOpt;
+
     cacheLoader_.load(latestSequence);
     latestSequence++;
 
@@ -217,7 +238,7 @@ ETLService::monitorReadOnly()
 void
 ETLService::run()
 {
-    log_.info() << "Starting reporting etl";
+    LOG(log_.info()) << "Starting reporting etl";
     state_.isStopping = false;
 
     doWork();
@@ -227,7 +248,7 @@ void
 ETLService::doWork()
 {
     worker_ = std::thread([this]() {
-        beast::setCurrentThreadName("rippled: ETLService worker");
+        beast::setCurrentThreadName("ETLService worker");
 
         if (state_.isReadOnly)
             monitorReadOnly();
@@ -237,7 +258,7 @@ ETLService::doWork()
 }
 
 ETLService::ETLService(
-    clio::Config const& config,
+    util::Config const& config,
     boost::asio::io_context& ioc,
     std::shared_ptr<BackendInterface> backend,
     std::shared_ptr<SubscriptionManagerType> subscriptions,
@@ -250,6 +271,7 @@ ETLService::ETLService(
     , ledgerFetcher_(backend, balancer)
     , ledgerLoader_(backend, balancer, ledgerFetcher_, state_)
     , ledgerPublisher_(ioc, backend, subscriptions, state_)
+    , amendmentBlockHandler_(ioc, state_)
 {
     startSequence_ = config.maybeValue<uint32_t>("start_sequence");
     finishSequence_ = config.maybeValue<uint32_t>("finish_sequence");
@@ -257,3 +279,4 @@ ETLService::ETLService(
     extractorThreads_ = config.valueOr<uint32_t>("extractor_threads", extractorThreads_);
     txnThreshold_ = config.valueOr<size_t>("txn_threshold", txnThreshold_);
 }
+}  // namespace etl

src/etl/ETLService.h

@@ -19,11 +19,12 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
-#include <backend/LedgerCache.h>
+#include <data/BackendInterface.h>
+#include <data/LedgerCache.h>
 #include <etl/LoadBalancer.h>
 #include <etl/Source.h>
 #include <etl/SystemState.h>
+#include <etl/impl/AmendmentBlock.h>
 #include <etl/impl/CacheLoader.h>
 #include <etl/impl/ExtractionDataPipe.h>
 #include <etl/impl/Extractor.h>
@@ -31,10 +32,11 @@
 #include <etl/impl/LedgerLoader.h>
 #include <etl/impl/LedgerPublisher.h>
 #include <etl/impl/Transformer.h>
-#include <log/Logger.h>
-#include <subscriptions/SubscriptionManager.h>
+#include <feed/SubscriptionManager.h>
+#include <util/log/Logger.h>
 
-#include "org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h"
+#include <ripple/proto/org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h>
+#include <boost/asio/steady_timer.hpp>
 #include <grpcpp/grpcpp.h>
 
 #include <memory>
@@ -42,7 +44,14 @@
 struct AccountTransactionsData;
 struct NFTTransactionsData;
 struct NFTsData;
+namespace feed {
 class SubscriptionManager;
+}
+
+/**
+ * @brief This namespace contains everything to do with the ETL and ETL sources.
+ */
+namespace etl {
 
 /**
  * @brief This class is responsible for continuously extracting data from a p2p node, and writing that data to the
@@ -60,18 +69,20 @@ class SubscriptionManager;
 class ETLService
 {
     // TODO: make these template parameters in ETLService
-    using SubscriptionManagerType = SubscriptionManager;
+    using SubscriptionManagerType = feed::SubscriptionManager;
     using LoadBalancerType = LoadBalancer;
     using NetworkValidatedLedgersType = NetworkValidatedLedgers;
-    using DataPipeType = clio::detail::ExtractionDataPipe<org::xrpl::rpc::v1::GetLedgerResponse>;
-    using CacheLoaderType = clio::detail::CacheLoader<Backend::LedgerCache>;
-    using LedgerFetcherType = clio::detail::LedgerFetcher<LoadBalancerType>;
-    using ExtractorType = clio::detail::Extractor<DataPipeType, NetworkValidatedLedgersType, LedgerFetcherType>;
-    using LedgerLoaderType = clio::detail::LedgerLoader<LoadBalancerType, LedgerFetcherType>;
-    using LedgerPublisherType = clio::detail::LedgerPublisher<SubscriptionManagerType>;
-    using TransformerType = clio::detail::Transformer<DataPipeType, LedgerLoaderType, LedgerPublisherType>;
+    using DataPipeType = etl::detail::ExtractionDataPipe<org::xrpl::rpc::v1::GetLedgerResponse>;
+    using CacheLoaderType = etl::detail::CacheLoader<data::LedgerCache>;
+    using LedgerFetcherType = etl::detail::LedgerFetcher<LoadBalancerType>;
+    using ExtractorType = etl::detail::Extractor<DataPipeType, NetworkValidatedLedgersType, LedgerFetcherType>;
+    using LedgerLoaderType = etl::detail::LedgerLoader<LoadBalancerType, LedgerFetcherType>;
+    using LedgerPublisherType = etl::detail::LedgerPublisher<SubscriptionManagerType>;
+    using AmendmentBlockHandlerType = etl::detail::AmendmentBlockHandler<>;
+    using TransformerType =
+        etl::detail::Transformer<DataPipeType, LedgerLoaderType, LedgerPublisherType, AmendmentBlockHandlerType>;
 
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::shared_ptr<BackendInterface> backend_;
     std::shared_ptr<LoadBalancerType> loadBalancer_;
@@ -84,6 +95,7 @@ class ETLService
     LedgerFetcherType ledgerFetcher_;
     LedgerLoaderType ledgerLoader_;
     LedgerPublisherType ledgerPublisher_;
+    AmendmentBlockHandlerType amendmentBlockHandler_;
 
     SystemState state_;
 
@@ -94,7 +106,7 @@ class ETLService
 
 public:
     /**
-     * @brief Create an instance of ETLService
+     * @brief Create an instance of ETLService.
      *
      * @param config The configuration to use
      * @param ioc io context to run on
@@ -104,16 +116,28 @@ public:
      * @param ledgers The network validated ledgers datastructure
      */
     ETLService(
-        clio::Config const& config,
+        util::Config const& config,
         boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> backend,
         std::shared_ptr<SubscriptionManagerType> subscriptions,
         std::shared_ptr<LoadBalancerType> balancer,
         std::shared_ptr<NetworkValidatedLedgersType> ledgers);
 
+    /**
+     * @brief A factory function to spawn new ETLService instances.
+     *
+     * Creates and runs the ETL service.
+     *
+     * @param config The configuration to use
+     * @param ioc io context to run on
+     * @param backend BackendInterface implementation
+     * @param subscriptions Subscription manager
+     * @param balancer Load balancer to use
+     * @param ledgers The network validated ledgers datastructure
+     */
     static std::shared_ptr<ETLService>
     make_ETLService(
-        clio::Config const& config,
+        util::Config const& config,
         boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> backend,
         std::shared_ptr<SubscriptionManagerType> subscriptions,
@@ -127,12 +151,12 @@ public:
     }
 
     /**
-     * @brief Stops components and joins worker thread
+     * @brief Stops components and joins worker thread.
      */
     ~ETLService()
     {
-        log_.info() << "onStop called";
-        log_.debug() << "Stopping Reporting ETL";
+        LOG(log_.info()) << "onStop called";
+        LOG(log_.debug()) << "Stopping Reporting ETL";
 
         state_.isStopping = true;
         cacheLoader_.stop();
@@ -140,11 +164,11 @@ public:
         if (worker_.joinable())
             worker_.join();
 
-        log_.debug() << "Joined ETLService worker thread";
+        LOG(log_.debug()) << "Joined ETLService worker thread";
     }
 
     /**
-     * @brief Get time passed since last ledger close, in seconds
+     * @brief Get time passed since last ledger close, in seconds.
      */
     std::uint32_t
     lastCloseAgeSeconds() const
@@ -152,6 +176,17 @@ public:
         return ledgerPublisher_.lastCloseAgeSeconds();
     }
 
+    /**
+     * @brief Check for the amendment blocked state.
+     *
+     * @return true if currently amendment blocked; false otherwise
+     */
+    bool
+    isAmendmentBlocked() const
+    {
+        return state_.isAmendmentBlocked;
+    }
+
     /**
      * @brief Get state of ETL as a JSON object
      */
@@ -177,10 +212,11 @@ private:
      * @note database must already be populated when this function is called
      *
      * @param startSequence the first ledger to extract
+     * @param numExtractors number of extractors to use
      * @return the last ledger written to the database, if any
      */
     std::optional<uint32_t>
-    runETLPipeline(uint32_t startSequence, int offset);
+    runETLPipeline(uint32_t startSequence, uint32_t numExtractors);
 
     /**
      * @brief Monitor the network for newly validated ledgers.
@@ -194,6 +230,15 @@ private:
     void
     monitor();
 
+    /**
+     * @brief Monitor the network for newly validated ledgers and publish them to the ledgers stream
+     *
+     * @param nextSequence the ledger sequence to publish
+     * @return the next ledger sequence to publish
+     */
+    uint32_t
+    publishNextSequence(uint32_t nextSequence);
+
     /**
      * @brief Monitor the database for newly written ledgers.
      *
@@ -213,7 +258,7 @@ private:
     }
 
     /**
-     * @brief Get the number of markers to use during the initial ledger download
+     * @brief Get the number of markers to use during the initial ledger download.
      *
      * This is equivelent to the degree of parallelism during the initial ledger download.
      *
@@ -226,14 +271,15 @@ private:
     }
 
     /**
-     * @brief Start all components to run ETL service
+     * @brief Start all components to run ETL service.
      */
     void
     run();
 
     /**
-     * @brief Spawn the worker thread and start monitoring
+     * @brief Spawn the worker thread and start monitoring.
      */
     void
     doWork();
 };
+}  // namespace etl

src/etl/LoadBalancer.cpp

@@ -17,14 +17,14 @@
 */
 //==============================================================================
 
-#include <backend/DBHelpers.h>
+#include <data/DBHelpers.h>
 #include <etl/ETLService.h>
 #include <etl/NFTHelpers.h>
 #include <etl/ProbingSource.h>
 #include <etl/Source.h>
-#include <log/Logger.h>
 #include <rpc/RPCHelpers.h>
 #include <util/Profiler.h>
+#include <util/log/Logger.h>
 
 #include <ripple/beast/net/IPEndpoint.h>
 #include <ripple/protocol/STLedgerEntry.h>
@@ -36,20 +36,20 @@
 
 #include <thread>
 
-using namespace clio;
+using namespace util;
+
+namespace etl {
 
 std::unique_ptr<Source>
 LoadBalancer::make_Source(
-    clio::Config const& config,
-    boost::asio::io_context& ioContext,
+    Config const& config,
+    boost::asio::io_context& ioc,
     std::shared_ptr<BackendInterface> backend,
-    std::shared_ptr<SubscriptionManager> subscriptions,
-    std::shared_ptr<NetworkValidatedLedgers> networkValidatedLedgers,
+    std::shared_ptr<feed::SubscriptionManager> subscriptions,
+    std::shared_ptr<NetworkValidatedLedgers> validatedLedgers,
     LoadBalancer& balancer)
 {
-    auto src =
-        std::make_unique<ProbingSource>(config, ioContext, backend, subscriptions, networkValidatedLedgers, balancer);
-
+    auto src = std::make_unique<ProbingSource>(config, ioc, backend, subscriptions, validatedLedgers, balancer);
     src->run();
 
     return src;
@@ -57,21 +57,21 @@ LoadBalancer::make_Source(
 
 std::shared_ptr<LoadBalancer>
 LoadBalancer::make_LoadBalancer(
-    clio::Config const& config,
+    Config const& config,
     boost::asio::io_context& ioc,
     std::shared_ptr<BackendInterface> backend,
-    std::shared_ptr<SubscriptionManager> subscriptions,
+    std::shared_ptr<feed::SubscriptionManager> subscriptions,
     std::shared_ptr<NetworkValidatedLedgers> validatedLedgers)
 {
     return std::make_shared<LoadBalancer>(config, ioc, backend, subscriptions, validatedLedgers);
 }
 
 LoadBalancer::LoadBalancer(
-    clio::Config const& config,
-    boost::asio::io_context& ioContext,
+    Config const& config,
+    boost::asio::io_context& ioc,
     std::shared_ptr<BackendInterface> backend,
-    std::shared_ptr<SubscriptionManager> subscriptions,
-    std::shared_ptr<NetworkValidatedLedgers> nwvl)
+    std::shared_ptr<feed::SubscriptionManager> subscriptions,
+    std::shared_ptr<NetworkValidatedLedgers> validatedLedgers)
 {
     if (auto value = config.maybeValue<uint32_t>("num_markers"); value)
         downloadRanges_ = std::clamp(*value, 1u, 256u);
@@ -80,13 +80,18 @@ LoadBalancer::LoadBalancer(
 
     for (auto const& entry : config.array("etl_sources"))
     {
-        std::unique_ptr<Source> source = make_Source(entry, ioContext, backend, subscriptions, nwvl, *this);
+        std::unique_ptr<Source> source = make_Source(entry, ioc, backend, subscriptions, validatedLedgers, *this);
 
         sources_.push_back(std::move(source));
-        log_.info() << "Added etl source - " << sources_.back()->toString();
+        LOG(log_.info()) << "Added etl source - " << sources_.back()->toString();
     }
 }
 
+LoadBalancer::~LoadBalancer()
+{
+    sources_.clear();
+}
+
 std::pair<std::vector<std::string>, bool>
 LoadBalancer::loadInitialLedger(uint32_t sequence, bool cacheOnly)
 {
@@ -96,8 +101,8 @@ LoadBalancer::loadInitialLedger(uint32_t sequence, bool cacheOnly)
             auto [data, res] = source->loadInitialLedger(sequence, downloadRanges_, cacheOnly);
 
             if (!res)
-                log_.error() << "Failed to download initial ledger."
-                             << " Sequence = " << sequence << " source = " << source->toString();
+                LOG(log_.error()) << "Failed to download initial ledger."
+                                  << " Sequence = " << sequence << " source = " << source->toString();
             else
                 response = std::move(data);
 
@@ -117,15 +122,15 @@ LoadBalancer::fetchLedger(uint32_t ledgerSequence, bool getObjects, bool getObje
             response = std::move(data);
             if (status.ok() && response.validated())
             {
-                log.info() << "Successfully fetched ledger = " << ledgerSequence
-                           << " from source = " << source->toString();
+                LOG(log.info()) << "Successfully fetched ledger = " << ledgerSequence
+                                << " from source = " << source->toString();
                 return true;
             }
             else
             {
-                log.warn() << "Could not fetch ledger " << ledgerSequence << ", Reply: " << response.DebugString()
-                           << ", error_code: " << status.error_code() << ", error_msg: " << status.error_message()
-                           << ", source = " << source->toString();
+                LOG(log.warn()) << "Could not fetch ledger " << ledgerSequence << ", Reply: " << response.DebugString()
+                                << ", error_code: " << status.error_code() << ", error_msg: " << status.error_message()
+                                << ", source = " << source->toString();
                 return false;
             }
         },
@@ -140,11 +145,11 @@ std::optional<boost::json::object>
 LoadBalancer::forwardToRippled(
     boost::json::object const& request,
     std::string const& clientIp,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
-    srand((unsigned)time(0));
+    srand(static_cast<unsigned>(time(0)));
     auto sourceIdx = rand() % sources_.size();
-    auto numAttempts = 0;
+    auto numAttempts = 0u;
 
     while (numAttempts < sources_.size())
     {
@@ -188,7 +193,7 @@ template <class Func>
 bool
 LoadBalancer::execute(Func f, uint32_t ledgerSequence)
 {
-    srand((unsigned)time(0));
+    srand(static_cast<unsigned>(time(0)));
     auto sourceIdx = rand() % sources_.size();
     auto numAttempts = 0;
 
@@ -196,8 +201,8 @@ LoadBalancer::execute(Func f, uint32_t ledgerSequence)
     {
         auto& source = sources_[sourceIdx];
 
-        log_.debug() << "Attempting to execute func. ledger sequence = " << ledgerSequence
-                     << " - source = " << source->toString();
+        LOG(log_.debug()) << "Attempting to execute func. ledger sequence = " << ledgerSequence
+                          << " - source = " << source->toString();
         // Originally, it was (source->hasLedger(ledgerSequence) || true)
         /* Sometimes rippled has ledger but doesn't actually know. However,
         but this does NOT happen in the normal case and is safe to remove
@@ -207,29 +212,31 @@ LoadBalancer::execute(Func f, uint32_t ledgerSequence)
             bool res = f(source);
             if (res)
             {
-                log_.debug() << "Successfully executed func at source = " << source->toString()
-                             << " - ledger sequence = " << ledgerSequence;
+                LOG(log_.debug()) << "Successfully executed func at source = " << source->toString()
+                                  << " - ledger sequence = " << ledgerSequence;
                 break;
             }
             else
             {
-                log_.warn() << "Failed to execute func at source = " << source->toString()
-                            << " - ledger sequence = " << ledgerSequence;
+                LOG(log_.warn()) << "Failed to execute func at source = " << source->toString()
+                                 << " - ledger sequence = " << ledgerSequence;
             }
         }
         else
         {
-            log_.warn() << "Ledger not present at source = " << source->toString()
-                        << " - ledger sequence = " << ledgerSequence;
+            LOG(log_.warn()) << "Ledger not present at source = " << source->toString()
+                             << " - ledger sequence = " << ledgerSequence;
         }
         sourceIdx = (sourceIdx + 1) % sources_.size();
         numAttempts++;
         if (numAttempts % sources_.size() == 0)
         {
-            log_.info() << "Ledger sequence " << ledgerSequence << " is not yet available from any configured sources. "
-                        << "Sleeping and trying again";
+            LOG(log_.info()) << "Ledger sequence " << ledgerSequence
+                             << " is not yet available from any configured sources. "
+                             << "Sleeping and trying again";
             std::this_thread::sleep_for(std::chrono::seconds(2));
         }
     }
     return true;
 }
+}  // namespace etl

src/etl/LoadBalancer.h

@@ -19,22 +19,29 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
-#include <config/Config.h>
+#include <data/BackendInterface.h>
 #include <etl/ETLHelpers.h>
-#include <log/Logger.h>
-#include <subscriptions/SubscriptionManager.h>
+#include <feed/SubscriptionManager.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
 
+#include <ripple/proto/org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h>
 #include <boost/asio.hpp>
-#include "org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h"
 #include <grpcpp/grpcpp.h>
 
+namespace etl {
 class Source;
 class ProbingSource;
+}  // namespace etl
+
+namespace feed {
 class SubscriptionManager;
+}  // namespace feed
+
+namespace etl {
 
 /**
- * @brief This class is used to manage connections to transaction processing processes
+ * @brief This class is used to manage connections to transaction processing processes.
  *
  * This class spawns a listener for each etl source, which listens to messages on the ledgers stream (to keep track of
  * which ledgers have been validated by the network, and the range of ledgers each etl source has). This class also
@@ -48,66 +55,84 @@ public:
     using OptionalGetLedgerResponseType = std::optional<GetLedgerResponseType>;
 
 private:
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
     std::vector<std::unique_ptr<Source>> sources_;
-    std::uint32_t downloadRanges_ = 16;
+    std::uint32_t downloadRanges_ = 16; /*< The number of markers to use when downloading intial ledger */
 
 public:
     /**
-     * @brief Create an instance of the load balancer
+     * @brief Create an instance of the load balancer.
      *
      * @param config The configuration to use
-     * @param ioContext io context to run on
+     * @param ioc The io_context to run on
      * @param backend BackendInterface implementation
      * @param subscriptions Subscription manager
-     * @param nwvl The network validated ledgers datastructure
+     * @param validatedLedgers The network validated ledgers datastructure
      */
     LoadBalancer(
-        clio::Config const& config,
-        boost::asio::io_context& ioContext,
-        std::shared_ptr<BackendInterface> backend,
-        std::shared_ptr<SubscriptionManager> subscriptions,
-        std::shared_ptr<NetworkValidatedLedgers> nwvl);
-
-    static std::shared_ptr<LoadBalancer>
-    make_LoadBalancer(
-        clio::Config const& config,
+        util::Config const& config,
         boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> backend,
-        std::shared_ptr<SubscriptionManager> subscriptions,
+        std::shared_ptr<feed::SubscriptionManager> subscriptions,
         std::shared_ptr<NetworkValidatedLedgers> validatedLedgers);
 
-    static std::unique_ptr<Source>
-    make_Source(
-        clio::Config const& config,
-        boost::asio::io_context& ioContext,
+    /**
+     * @brief A factory function for the load balancer.
+     *
+     * @param config The configuration to use
+     * @param ioc The io_context to run on
+     * @param backend BackendInterface implementation
+     * @param subscriptions Subscription manager
+     * @param validatedLedgers The network validated ledgers datastructure
+     */
+    static std::shared_ptr<LoadBalancer>
+    make_LoadBalancer(
+        util::Config const& config,
+        boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> backend,
-        std::shared_ptr<SubscriptionManager> subscriptions,
-        std::shared_ptr<NetworkValidatedLedgers> networkValidatedLedgers,
-        LoadBalancer& balancer);
-
-    ~LoadBalancer()
-    {
-        sources_.clear();
-    }
+        std::shared_ptr<feed::SubscriptionManager> subscriptions,
+        std::shared_ptr<NetworkValidatedLedgers> validatedLedgers);
 
     /**
-     * @brief Load the initial ledger, writing data to the queue
+     * @brief A factory function for the ETL source.
      *
-     * @param sequence sequence of ledger to download
+     * @param config The configuration to use
+     * @param ioc The io_context to run on
+     * @param backend BackendInterface implementation
+     * @param subscriptions Subscription manager
+     * @param validatedLedgers The network validated ledgers datastructure
+     * @param balancer The load balancer
+     */
+    static std::unique_ptr<Source>
+    make_Source(
+        util::Config const& config,
+        boost::asio::io_context& ioc,
+        std::shared_ptr<BackendInterface> backend,
+        std::shared_ptr<feed::SubscriptionManager> subscriptions,
+        std::shared_ptr<NetworkValidatedLedgers> validatedLedgers,
+        LoadBalancer& balancer);
+
+    ~LoadBalancer();
+
+    /**
+     * @brief Load the initial ledger, writing data to the queue.
+     *
+     * @param sequence Sequence of ledger to download
+     * @param cacheOnly Whether to only write to cache and not to the DB; defaults to false
      */
     std::pair<std::vector<std::string>, bool>
     loadInitialLedger(uint32_t sequence, bool cacheOnly = false);
 
     /**
-     * @brief Fetch data for a specific ledger
+     * @brief Fetch data for a specific ledger.
      *
      * This function will continuously try to fetch data for the specified ledger until the fetch succeeds, the ledger
      * is found in the database, or the server is shutting down.
      *
-     * @param ledgerSequence sequence of ledger to fetch data for
-     * @param getObjects if true, fetch diff between specified ledger and previous
-     * @return the extracted data, if extraction was successful. If the ledger was found in the database or the server
+     * @param ledgerSequence Sequence of the ledger to fetch
+     * @param getObjects Whether to get the account state diff between this ledger and the prior one
+     * @param getObjectNeighbors Whether to request object neighbors
+     * @return The extracted data, if extraction was successful. If the ledger was found in the database or the server
      * is shutting down, the optional will be empty
      */
     OptionalGetLedgerResponseType
@@ -127,30 +152,32 @@ public:
     shouldPropagateTxnStream(Source* in) const;
 
     /**
-     * @return JSON representation of the state of this load balancer
+     * @return JSON representation of the state of this load balancer.
      */
     boost::json::value
     toJson() const;
 
     /**
-     * @brief Forward a JSON RPC request to a randomly selected rippled node
+     * @brief Forward a JSON RPC request to a randomly selected rippled node.
      *
-     * @param request JSON-RPC request
-     * @return response received from rippled node
+     * @param request JSON-RPC request to forward
+     * @param clientIp The IP address of the peer
+     * @param yield The coroutine context
+     * @return Response received from rippled node as JSON object on success; nullopt on failure
      */
     std::optional<boost::json::object>
-    forwardToRippled(boost::json::object const& request, std::string const& clientIp, boost::asio::yield_context& yield)
+    forwardToRippled(boost::json::object const& request, std::string const& clientIp, boost::asio::yield_context yield)
         const;
 
 private:
     /**
-     * @brief Execute a function on a randomly selected source
+     * @brief Execute a function on a randomly selected source.
      *
      * @note f is a function that takes an Source as an argument and returns a bool.
      * Attempt to execute f for one randomly chosen Source that has the specified ledger. If f returns false, another
      * randomly chosen Source is used. The process repeats until f returns true.
      *
-     * @param f function to execute. This function takes the ETL source as an argument, and returns a bool.
+     * @param f Function to execute. This function takes the ETL source as an argument, and returns a bool
      * @param ledgerSequence f is executed for each Source that has this ledger
      * @return true if f was eventually executed successfully. false if the ledger was found in the database or the
      * server is shutting down
@@ -159,3 +186,4 @@ private:
     bool
     execute(Func f, uint32_t ledgerSequence);
 };
+}  // namespace etl

src/etl/NFTHelpers.cpp

@@ -22,9 +22,12 @@
 #include <ripple/protocol/TxMeta.h>
 #include <vector>
 
-#include <backend/BackendInterface.h>
-#include <backend/DBHelpers.h>
-#include <backend/Types.h>
+#include <data/BackendInterface.h>
+#include <data/DBHelpers.h>
+#include <data/Types.h>
+#include <fmt/core.h>
+
+namespace etl {
 
 std::pair<std::vector<NFTTransactionsData>, std::optional<NFTsData>>
 getNFTokenMintData(ripple::TxMeta const& txMeta, ripple::STTx const& sttx)
@@ -95,21 +98,20 @@ getNFTokenMintData(ripple::TxMeta const& txMeta, ripple::STTx const& sttx)
 
     std::sort(finalIDs.begin(), finalIDs.end());
     std::sort(prevIDs.begin(), prevIDs.end());
-    std::vector<ripple::uint256> tokenIDResult;
-    std::set_difference(
-        finalIDs.begin(),
-        finalIDs.end(),
-        prevIDs.begin(),
-        prevIDs.end(),
-        std::inserter(tokenIDResult, tokenIDResult.begin()));
-    if (tokenIDResult.size() == 1 && owner)
-        return {
-            {NFTTransactionsData(tokenIDResult.front(), txMeta, sttx.getTransactionID())},
-            NFTsData(tokenIDResult.front(), *owner, sttx.getFieldVL(ripple::sfURI), txMeta)};
 
-    std::stringstream msg;
-    msg << " - unexpected NFTokenMint data in tx " << sttx.getTransactionID();
-    throw std::runtime_error(msg.str());
+    // Find the first NFT ID that doesn't match.  We're looking for an
+    // added NFT, so the one we want will be the mismatch in finalIDs.
+    auto const diff = std::mismatch(finalIDs.begin(), finalIDs.end(), prevIDs.begin(), prevIDs.end());
+
+    // There should always be a difference so the returned finalIDs
+    // iterator should never be end().  But better safe than sorry.
+    if (finalIDs.size() != prevIDs.size() + 1 || diff.first == finalIDs.end() || !owner)
+        throw std::runtime_error(
+            fmt::format(" - unexpected NFTokenMint data in tx {}", strHex(sttx.getTransactionID())));
+
+    return {
+        {NFTTransactionsData(*diff.first, txMeta, sttx.getTransactionID())},
+        NFTsData(*diff.first, *owner, sttx.getFieldVL(ripple::sfURI), txMeta)};
 }
 
 std::pair<std::vector<NFTTransactionsData>, std::optional<NFTsData>>
@@ -338,3 +340,4 @@ getNFTDataFromObj(std::uint32_t const seq, std::string const& key, std::string c
 
     return nfts;
 }
+}  // namespace etl

src/etl/NFTHelpers.h

@@ -17,21 +17,35 @@
 */
 //==============================================================================
 
+/** @file */
 #pragma once
 
-#include <backend/DBHelpers.h>
+#include <data/DBHelpers.h>
 
 #include <ripple/protocol/STTx.h>
 #include <ripple/protocol/TxMeta.h>
 
+namespace etl {
+
 /**
- * @brief Pull NFT data from TX via ETLService
+ * @brief Pull NFT data from TX via ETLService.
+ *
+ * @param txMeta Transaction metadata
+ * @param sttx The transaction
+ * @return NFT transactions data as a pair of transactions and optional NFTsData
  */
 std::pair<std::vector<NFTTransactionsData>, std::optional<NFTsData>>
 getNFTDataFromTx(ripple::TxMeta const& txMeta, ripple::STTx const& sttx);
 
 /**
- * @brief Pull NFT data from ledger object via loadInitialLedger
+ * @brief Pull NFT data from ledger object via loadInitialLedger.
+ *
+ * @param seq The ledger sequence to pull for
+ * @param key The owner key
+ * @param blob Object data as blob
+ * @return The NFT data as a vector
  */
 std::vector<NFTsData>
 getNFTDataFromObj(std::uint32_t const seq, std::string const& key, std::string const& blob);
+
+}  // namespace etl

src/etl/ProbingSource.cpp

@@ -18,15 +18,14 @@
 //==============================================================================
 
 #include <etl/ProbingSource.h>
-#include <log/Logger.h>
 
-using namespace clio;
+namespace etl {
 
 ProbingSource::ProbingSource(
-    clio::Config const& config,
+    util::Config const& config,
     boost::asio::io_context& ioc,
     std::shared_ptr<BackendInterface> backend,
-    std::shared_ptr<SubscriptionManager> subscriptions,
+    std::shared_ptr<feed::SubscriptionManager> subscriptions,
     std::shared_ptr<NetworkValidatedLedgers> nwvl,
     LoadBalancer& balancer,
     boost::asio::ssl::context sslCtx)
@@ -106,26 +105,26 @@ ProbingSource::token() const
 }
 
 std::pair<std::vector<std::string>, bool>
-ProbingSource::loadInitialLedger(std::uint32_t ledgerSequence, std::uint32_t numMarkers, bool cacheOnly)
+ProbingSource::loadInitialLedger(std::uint32_t sequence, std::uint32_t numMarkers, bool cacheOnly)
 {
     if (!currentSrc_)
         return {{}, false};
-    return currentSrc_->loadInitialLedger(ledgerSequence, numMarkers, cacheOnly);
+    return currentSrc_->loadInitialLedger(sequence, numMarkers, cacheOnly);
 }
 
 std::pair<grpc::Status, ProbingSource::GetLedgerResponseType>
-ProbingSource::fetchLedger(uint32_t ledgerSequence, bool getObjects, bool getObjectNeighbors)
+ProbingSource::fetchLedger(uint32_t sequence, bool getObjects, bool getObjectNeighbors)
 {
     if (!currentSrc_)
         return {};
-    return currentSrc_->fetchLedger(ledgerSequence, getObjects, getObjectNeighbors);
+    return currentSrc_->fetchLedger(sequence, getObjects, getObjectNeighbors);
 }
 
 std::optional<boost::json::object>
 ProbingSource::forwardToRippled(
     boost::json::object const& request,
     std::string const& clientIp,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     if (!currentSrc_)
         return {};
@@ -136,7 +135,7 @@ std::optional<boost::json::object>
 ProbingSource::requestFromRippled(
     boost::json::object const& request,
     std::string const& clientIp,
-    boost::asio::yield_context& yield) const
+    boost::asio::yield_context yield) const
 {
     if (!currentSrc_)
         return {};
@@ -156,12 +155,12 @@ ProbingSource::make_SSLHooks() noexcept
                 {
                     plainSrc_->pause();
                     currentSrc_ = sslSrc_;
-                    log_.info() << "Selected WSS as the main source: " << currentSrc_->toString();
+                    LOG(log_.info()) << "Selected WSS as the main source: " << currentSrc_->toString();
                 }
                 return SourceHooks::Action::PROCEED;
             },
             // onDisconnected
-            [this](auto ec) {
+            [this](auto /* ec */) {
                 std::lock_guard lck(mtx_);
                 if (currentSrc_)
                 {
@@ -185,12 +184,12 @@ ProbingSource::make_PlainHooks() noexcept
                 {
                     sslSrc_->pause();
                     currentSrc_ = plainSrc_;
-                    log_.info() << "Selected Plain WS as the main source: " << currentSrc_->toString();
+                    LOG(log_.info()) << "Selected Plain WS as the main source: " << currentSrc_->toString();
                 }
                 return SourceHooks::Action::PROCEED;
             },
             // onDisconnected
-            [this](auto ec) {
+            [this](auto /* ec */) {
                 std::lock_guard lck(mtx_);
                 if (currentSrc_)
                 {
@@ -199,4 +198,5 @@ ProbingSource::make_PlainHooks() noexcept
                 }
                 return SourceHooks::Action::STOP;
             }};
-}
+};
+}  // namespace etl

src/etl/ProbingSource.h

@@ -19,9 +19,9 @@
 
 #pragma once
 
-#include <config/Config.h>
 #include <etl/Source.h>
-#include <log/Logger.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
 
 #include <boost/asio.hpp>
 #include <boost/beast/core.hpp>
@@ -31,6 +31,8 @@
 
 #include <mutex>
 
+namespace etl {
+
 /**
  * @brief This Source implementation attempts to connect over both secure websocket and plain websocket.
  *
@@ -44,7 +46,7 @@ public:
     using GetLedgerResponseType = org::xrpl::rpc::v1::GetLedgerResponse;
 
 private:
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::mutex mtx_;
     boost::asio::ssl::context sslCtx_;
@@ -54,7 +56,7 @@ private:
 
 public:
     /**
-     * @brief Create an instance of the probing source
+     * @brief Create an instance of the probing source.
      *
      * @param config The configuration to use
      * @param ioc io context to run on
@@ -65,10 +67,10 @@ public:
      * @param sslCtx The SSL context to use; defaults to tlsv12
      */
     ProbingSource(
-        clio::Config const& config,
+        util::Config const& config,
         boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> backend,
-        std::shared_ptr<SubscriptionManager> subscriptions,
+        std::shared_ptr<feed::SubscriptionManager> subscriptions,
         std::shared_ptr<NetworkValidatedLedgers> nwvl,
         LoadBalancer& balancer,
         boost::asio::ssl::context sslCtx = boost::asio::ssl::context{boost::asio::ssl::context::tlsv12});
@@ -97,13 +99,13 @@ public:
     toString() const override;
 
     std::pair<std::vector<std::string>, bool>
-    loadInitialLedger(std::uint32_t ledgerSequence, std::uint32_t numMarkers, bool cacheOnly = false) override;
+    loadInitialLedger(std::uint32_t sequence, std::uint32_t numMarkers, bool cacheOnly = false) override;
 
     std::pair<grpc::Status, GetLedgerResponseType>
-    fetchLedger(uint32_t ledgerSequence, bool getObjects = true, bool getObjectNeighbors = false) override;
+    fetchLedger(uint32_t sequence, bool getObjects = true, bool getObjectNeighbors = false) override;
 
     std::optional<boost::json::object>
-    forwardToRippled(boost::json::object const& request, std::string const& clientIp, boost::asio::yield_context& yield)
+    forwardToRippled(boost::json::object const& request, std::string const& clientIp, boost::asio::yield_context yield)
         const override;
 
     boost::uuids::uuid
@@ -114,7 +116,7 @@ private:
     requestFromRippled(
         boost::json::object const& request,
         std::string const& clientIp,
-        boost::asio::yield_context& yield) const override;
+        boost::asio::yield_context yield) const override;
 
     SourceHooks
     make_SSLHooks() noexcept;
@@ -122,3 +124,4 @@ private:
     SourceHooks
     make_PlainHooks() noexcept;
 };
+}  // namespace etl

src/etl/README.md

@@ -1,3 +1,5 @@
+# ETL subsystem
+
 A single clio node has one or more ETL sources, specified in the config
 file. clio will subscribe to the `ledgers` stream of each of the ETL
 sources. This stream sends a message whenever a new ledger is validated. Upon

src/etl/Source.cpp

@@ -17,13 +17,11 @@
 */
 //==============================================================================
 
-#include <backend/DBHelpers.h>
+#include <data/DBHelpers.h>
 #include <etl/ETLService.h>
 #include <etl/LoadBalancer.h>
-#include <etl/NFTHelpers.h>
 #include <etl/ProbingSource.h>
 #include <etl/Source.h>
-#include <log/Logger.h>
 #include <rpc/RPCHelpers.h>
 #include <util/Profiler.h>
 
@@ -36,82 +34,19 @@
 
 #include <thread>
 
-using namespace clio;
+namespace etl {
 
 static boost::beast::websocket::stream_base::timeout
 make_TimeoutOption()
 {
-    // See #289 for details.
-    // TODO: investigate the issue and find if there is a solution other than
-    // introducing artificial timeouts.
-    if (true)
-    {
-        // The only difference between this and the suggested client role is
-        // that idle_timeout is set to 20 instead of none()
-        auto opt = boost::beast::websocket::stream_base::timeout{};
-        opt.handshake_timeout = std::chrono::seconds(30);
-        opt.idle_timeout = std::chrono::seconds(20);
-        opt.keep_alive_pings = false;
-        return opt;
-    }
-    else
-    {
-        return boost::beast::websocket::stream_base::timeout::suggested(boost::beast::role_type::client);
-    }
-}
-
-template <class Derived>
-void
-SourceImpl<Derived>::reconnect(boost::beast::error_code ec)
-{
-    if (paused_)
-        return;
-
-    if (connected_)
-        hooks_.onDisconnected(ec);
-
-    connected_ = false;
-    // These are somewhat normal errors. operation_aborted occurs on shutdown,
-    // when the timer is cancelled. connection_refused will occur repeatedly
-    std::string err = ec.message();
-    // if we cannot connect to the transaction processing process
-    if (ec.category() == boost::asio::error::get_ssl_category())
-    {
-        err = std::string(" (") + boost::lexical_cast<std::string>(ERR_GET_LIB(ec.value())) + "," +
-            boost::lexical_cast<std::string>(ERR_GET_REASON(ec.value())) + ") ";
-        // ERR_PACK /* crypto/err/err.h */
-        char buf[128];
-        ::ERR_error_string_n(ec.value(), buf, sizeof(buf));
-        err += buf;
-
-        std::cout << err << std::endl;
-    }
-
-    if (ec != boost::asio::error::operation_aborted && ec != boost::asio::error::connection_refused)
-    {
-        log_.error() << "error code = " << ec << " - " << toString();
-    }
-    else
-    {
-        log_.warn() << "error code = " << ec << " - " << toString();
-    }
-
-    // exponentially increasing timeouts, with a max of 30 seconds
-    size_t waitTime = std::min(pow(2, numFailures_), 30.0);
-    numFailures_++;
-    timer_.expires_after(boost::asio::chrono::seconds(waitTime));
-    timer_.async_wait([this](auto ec) {
-        bool startAgain = (ec != boost::asio::error::operation_aborted);
-        log_.trace() << "async_wait : ec = " << ec;
-        derived().close(startAgain);
-    });
+    return boost::beast::websocket::stream_base::timeout::suggested(boost::beast::role_type::client);
 }
 
 void
 PlainSource::close(bool startAgain)
 {
     timer_.cancel();
-    ioc_.post([this, startAgain]() {
+    boost::asio::post(strand_, [this, startAgain]() {
         if (closing_)
             return;
 
@@ -124,24 +59,19 @@ PlainSource::close(bool startAgain)
             derived().ws().async_close(boost::beast::websocket::close_code::normal, [this, startAgain](auto ec) {
                 if (ec)
                 {
-                    log_.error() << " async_close : "
-                                 << "error code = " << ec << " - " << toString();
+                    LOG(log_.error()) << "async_close: error code = " << ec << " - " << toString();
                 }
                 closing_ = false;
                 if (startAgain)
                 {
-                    ws_ = std::make_unique<boost::beast::websocket::stream<boost::beast::tcp_stream>>(
-                        boost::asio::make_strand(ioc_));
-
+                    ws_ = std::make_unique<StreamType>(strand_);
                     run();
                 }
             });
         }
         else if (startAgain)
         {
-            ws_ = std::make_unique<boost::beast::websocket::stream<boost::beast::tcp_stream>>(
-                boost::asio::make_strand(ioc_));
-
+            ws_ = std::make_unique<StreamType>(strand_);
             run();
         }
     });
@@ -151,68 +81,41 @@ void
 SslSource::close(bool startAgain)
 {
     timer_.cancel();
-    ioc_.post([this, startAgain]() {
+    boost::asio::post(strand_, [this, startAgain]() {
         if (closing_)
             return;
 
         if (derived().ws().is_open())
         {
-            // onStop() also calls close(). If the async_close is called twice,
-            // an assertion fails. Using closing_ makes sure async_close is only
-            // called once
+            // onStop() also calls close(). If the async_close is called twice, an assertion fails. Using closing_ makes
+            // sure async_close is only called once
             closing_ = true;
             derived().ws().async_close(boost::beast::websocket::close_code::normal, [this, startAgain](auto ec) {
                 if (ec)
                 {
-                    log_.error() << " async_close : "
-                                 << "error code = " << ec << " - " << toString();
+                    LOG(log_.error()) << "async_close: error code = " << ec << " - " << toString();
                 }
                 closing_ = false;
                 if (startAgain)
                 {
-                    ws_ = std::make_unique<
-                        boost::beast::websocket::stream<boost::beast::ssl_stream<boost::beast::tcp_stream>>>(
-                        boost::asio::make_strand(ioc_), *sslCtx_);
-
+                    ws_ = std::make_unique<StreamType>(strand_, *sslCtx_);
                     run();
                 }
             });
         }
         else if (startAgain)
         {
-            ws_ = std::make_unique<boost::beast::websocket::stream<boost::beast::ssl_stream<boost::beast::tcp_stream>>>(
-                boost::asio::make_strand(ioc_), *sslCtx_);
-
+            ws_ = std::make_unique<StreamType>(strand_, *sslCtx_);
             run();
         }
     });
 }
 
-template <class Derived>
-void
-SourceImpl<Derived>::onResolve(boost::beast::error_code ec, boost::asio::ip::tcp::resolver::results_type results)
-{
-    log_.trace() << "ec = " << ec << " - " << toString();
-    if (ec)
-    {
-        // try again
-        reconnect(ec);
-    }
-    else
-    {
-        boost::beast::get_lowest_layer(derived().ws()).expires_after(std::chrono::seconds(30));
-        boost::beast::get_lowest_layer(derived().ws()).async_connect(results, [this](auto ec, auto ep) {
-            derived().onConnect(ec, ep);
-        });
-    }
-}
-
 void
 PlainSource::onConnect(
     boost::beast::error_code ec,
     boost::asio::ip::tcp::resolver::results_type::endpoint_type endpoint)
 {
-    log_.trace() << "ec = " << ec << " - " << toString();
     if (ec)
     {
         // start over
@@ -220,19 +123,16 @@ PlainSource::onConnect(
     }
     else
     {
+        connected_ = true;
         numFailures_ = 0;
-        // Turn off timeout on the tcp stream, because websocket stream has it's
-        // own timeout system
+
+        // Websocket stream has it's own timeout system
         boost::beast::get_lowest_layer(derived().ws()).expires_never();
 
-        // Set a desired timeout for the websocket stream
         derived().ws().set_option(make_TimeoutOption());
-
-        // Set a decorator to change the User-Agent of the handshake
         derived().ws().set_option(
             boost::beast::websocket::stream_base::decorator([](boost::beast::websocket::request_type& req) {
                 req.set(boost::beast::http::field::user_agent, "clio-client");
-
                 req.set("X-User", "clio-client");
             }));
 
@@ -240,7 +140,6 @@ PlainSource::onConnect(
         // Host HTTP header during the WebSocket handshake.
         // See https://tools.ietf.org/html/rfc7230#section-5.4
         auto host = ip_ + ':' + std::to_string(endpoint.port());
-        // Perform the websocket handshake
         derived().ws().async_handshake(host, "/", [this](auto ec) { onHandshake(ec); });
     }
 }
@@ -248,7 +147,6 @@ PlainSource::onConnect(
 void
 SslSource::onConnect(boost::beast::error_code ec, boost::asio::ip::tcp::resolver::results_type::endpoint_type endpoint)
 {
-    log_.trace() << "ec = " << ec << " - " << toString();
     if (ec)
     {
         // start over
@@ -256,19 +154,16 @@ SslSource::onConnect(boost::beast::error_code ec, boost::asio::ip::tcp::resolver
     }
     else
     {
+        connected_ = true;
         numFailures_ = 0;
-        // Turn off timeout on the tcp stream, because websocket stream has it's
-        // own timeout system
+
+        // Websocket stream has it's own timeout system
         boost::beast::get_lowest_layer(derived().ws()).expires_never();
 
-        // Set a desired timeout for the websocket stream
         derived().ws().set_option(make_TimeoutOption());
-
-        // Set a decorator to change the User-Agent of the handshake
         derived().ws().set_option(
             boost::beast::websocket::stream_base::decorator([](boost::beast::websocket::request_type& req) {
                 req.set(boost::beast::http::field::user_agent, "clio-client");
-
                 req.set("X-User", "clio-client");
             }));
 
@@ -276,7 +171,6 @@ SslSource::onConnect(boost::beast::error_code ec, boost::asio::ip::tcp::resolver
         // Host HTTP header during the WebSocket handshake.
         // See https://tools.ietf.org/html/rfc7230#section-5.4
         auto host = ip_ + ':' + std::to_string(endpoint.port());
-        // Perform the websocket handshake
         ws().next_layer().async_handshake(
             boost::asio::ssl::stream_base::client, [this, endpoint](auto ec) { onSslHandshake(ec, endpoint); });
     }
@@ -293,527 +187,8 @@ SslSource::onSslHandshake(
     }
     else
     {
-        // Perform the websocket handshake
         auto host = ip_ + ':' + std::to_string(endpoint.port());
-        // Perform the websocket handshake
         ws().async_handshake(host, "/", [this](auto ec) { onHandshake(ec); });
     }
 }
-
-template <class Derived>
-void
-SourceImpl<Derived>::onHandshake(boost::beast::error_code ec)
-{
-    log_.trace() << "ec = " << ec << " - " << toString();
-    if (auto action = hooks_.onConnected(ec); action == SourceHooks::Action::STOP)
-        return;
-
-    if (ec)
-    {
-        // start over
-        reconnect(ec);
-    }
-    else
-    {
-        boost::json::object jv{
-            {"command", "subscribe"}, {"streams", {"ledger", "manifests", "validations", "transactions_proposed"}}};
-        std::string s = boost::json::serialize(jv);
-        log_.trace() << "Sending subscribe stream message";
-
-        derived().ws().set_option(
-            boost::beast::websocket::stream_base::decorator([](boost::beast::websocket::request_type& req) {
-                req.set(
-                    boost::beast::http::field::user_agent, std::string(BOOST_BEAST_VERSION_STRING) + " clio-client");
-
-                req.set("X-User", "coro-client");
-            }));
-
-        // Send the message
-        derived().ws().async_write(boost::asio::buffer(s), [this](auto ec, size_t size) { onWrite(ec, size); });
-    }
-}
-
-template <class Derived>
-void
-SourceImpl<Derived>::onWrite(boost::beast::error_code ec, size_t bytesWritten)
-{
-    log_.trace() << "ec = " << ec << " - " << toString();
-    if (ec)
-    {
-        // start over
-        reconnect(ec);
-    }
-    else
-    {
-        derived().ws().async_read(readBuffer_, [this](auto ec, size_t size) { onRead(ec, size); });
-    }
-}
-
-template <class Derived>
-void
-SourceImpl<Derived>::onRead(boost::beast::error_code ec, size_t size)
-{
-    log_.trace() << "ec = " << ec << " - " << toString();
-    // if error or error reading message, start over
-    if (ec)
-    {
-        reconnect(ec);
-    }
-    else
-    {
-        handleMessage();
-        boost::beast::flat_buffer buffer;
-        swap(readBuffer_, buffer);
-
-        log_.trace() << "calling async_read - " << toString();
-        derived().ws().async_read(readBuffer_, [this](auto ec, size_t size) { onRead(ec, size); });
-    }
-}
-
-template <class Derived>
-bool
-SourceImpl<Derived>::handleMessage()
-{
-    log_.trace() << toString();
-
-    setLastMsgTime();
-    connected_ = true;
-    try
-    {
-        std::string msg{static_cast<char const*>(readBuffer_.data().data()), readBuffer_.size()};
-        log_.trace() << msg;
-        boost::json::value raw = boost::json::parse(msg);
-        log_.trace() << "parsed";
-        boost::json::object response = raw.as_object();
-
-        uint32_t ledgerIndex = 0;
-        if (response.contains("result"))
-        {
-            boost::json::object result = response["result"].as_object();
-            if (result.contains("ledger_index"))
-            {
-                ledgerIndex = result["ledger_index"].as_int64();
-            }
-            if (result.contains("validated_ledgers"))
-            {
-                boost::json::string const& validatedLedgers = result["validated_ledgers"].as_string();
-
-                setValidatedRange({validatedLedgers.c_str(), validatedLedgers.size()});
-            }
-            log_.info() << "Received a message on ledger "
-                        << " subscription stream. Message : " << response << " - " << toString();
-        }
-        else if (response.contains("type") && response["type"] == "ledgerClosed")
-        {
-            log_.info() << "Received a message on ledger "
-                        << " subscription stream. Message : " << response << " - " << toString();
-            if (response.contains("ledger_index"))
-            {
-                ledgerIndex = response["ledger_index"].as_int64();
-            }
-            if (response.contains("validated_ledgers"))
-            {
-                boost::json::string const& validatedLedgers = response["validated_ledgers"].as_string();
-                setValidatedRange({validatedLedgers.c_str(), validatedLedgers.size()});
-            }
-        }
-        else
-        {
-            if (balancer_.shouldPropagateTxnStream(this))
-            {
-                if (response.contains("transaction"))
-                {
-                    forwardCache_.freshen();
-                    subscriptions_->forwardProposedTransaction(response);
-                }
-                else if (response.contains("type") && response["type"] == "validationReceived")
-                {
-                    subscriptions_->forwardValidation(response);
-                }
-                else if (response.contains("type") && response["type"] == "manifestReceived")
-                {
-                    subscriptions_->forwardManifest(response);
-                }
-            }
-        }
-
-        if (ledgerIndex != 0)
-        {
-            log_.trace() << "Pushing ledger sequence = " << ledgerIndex << " - " << toString();
-            networkValidatedLedgers_->push(ledgerIndex);
-        }
-        return true;
-    }
-    catch (std::exception const& e)
-    {
-        log_.error() << "Exception in handleMessage : " << e.what();
-        return false;
-    }
-}
-
-// TODO: move to detail
-class AsyncCallData
-{
-    clio::Logger log_{"ETL"};
-
-    std::unique_ptr<org::xrpl::rpc::v1::GetLedgerDataResponse> cur_;
-    std::unique_ptr<org::xrpl::rpc::v1::GetLedgerDataResponse> next_;
-
-    org::xrpl::rpc::v1::GetLedgerDataRequest request_;
-    std::unique_ptr<grpc::ClientContext> context_;
-
-    grpc::Status status_;
-    unsigned char nextPrefix_;
-
-    std::string lastKey_;
-
-public:
-    AsyncCallData(uint32_t seq, ripple::uint256 const& marker, std::optional<ripple::uint256> const& nextMarker)
-    {
-        request_.mutable_ledger()->set_sequence(seq);
-        if (marker.isNonZero())
-        {
-            request_.set_marker(marker.data(), marker.size());
-        }
-        request_.set_user("ETL");
-        nextPrefix_ = 0x00;
-        if (nextMarker)
-            nextPrefix_ = nextMarker->data()[0];
-
-        unsigned char prefix = marker.data()[0];
-
-        log_.debug() << "Setting up AsyncCallData. marker = " << ripple::strHex(marker)
-                     << " . prefix = " << ripple::strHex(std::string(1, prefix))
-                     << " . nextPrefix_ = " << ripple::strHex(std::string(1, nextPrefix_));
-
-        assert(nextPrefix_ > prefix || nextPrefix_ == 0x00);
-
-        cur_ = std::make_unique<org::xrpl::rpc::v1::GetLedgerDataResponse>();
-        next_ = std::make_unique<org::xrpl::rpc::v1::GetLedgerDataResponse>();
-        context_ = std::make_unique<grpc::ClientContext>();
-    }
-
-    enum class CallStatus { MORE, DONE, ERRORED };
-
-    CallStatus
-    process(
-        std::unique_ptr<org::xrpl::rpc::v1::XRPLedgerAPIService::Stub>& stub,
-        grpc::CompletionQueue& cq,
-        BackendInterface& backend,
-        bool abort,
-        bool cacheOnly = false)
-    {
-        log_.trace() << "Processing response. "
-                     << "Marker prefix = " << getMarkerPrefix();
-        if (abort)
-        {
-            log_.error() << "AsyncCallData aborted";
-            return CallStatus::ERRORED;
-        }
-        if (!status_.ok())
-        {
-            log_.error() << "AsyncCallData status_ not ok: "
-                         << " code = " << status_.error_code() << " message = " << status_.error_message();
-            return CallStatus::ERRORED;
-        }
-        if (!next_->is_unlimited())
-        {
-            log_.warn() << "AsyncCallData is_unlimited is false. Make sure "
-                           "secure_gateway is set correctly at the ETL source";
-        }
-
-        std::swap(cur_, next_);
-
-        bool more = true;
-
-        // if no marker returned, we are done
-        if (cur_->marker().size() == 0)
-            more = false;
-
-        // if returned marker is greater than our end, we are done
-        unsigned char prefix = cur_->marker()[0];
-        if (nextPrefix_ != 0x00 && prefix >= nextPrefix_)
-            more = false;
-
-        // if we are not done, make the next async call
-        if (more)
-        {
-            request_.set_marker(std::move(cur_->marker()));
-            call(stub, cq);
-        }
-
-        auto const numObjects = cur_->ledger_objects().objects_size();
-        log_.debug() << "Writing " << numObjects << " objects";
-
-        std::vector<Backend::LedgerObject> cacheUpdates;
-        cacheUpdates.reserve(numObjects);
-
-        for (int i = 0; i < numObjects; ++i)
-        {
-            auto& obj = *(cur_->mutable_ledger_objects()->mutable_objects(i));
-            if (!more && nextPrefix_ != 0x00)
-            {
-                if (((unsigned char)obj.key()[0]) >= nextPrefix_)
-                    continue;
-            }
-            cacheUpdates.push_back(
-                {*ripple::uint256::fromVoidChecked(obj.key()),
-                 {obj.mutable_data()->begin(), obj.mutable_data()->end()}});
-            if (!cacheOnly)
-            {
-                if (lastKey_.size())
-                    backend.writeSuccessor(std::move(lastKey_), request_.ledger().sequence(), std::string{obj.key()});
-                lastKey_ = obj.key();
-                backend.writeNFTs(getNFTDataFromObj(request_.ledger().sequence(), obj.key(), obj.data()));
-                backend.writeLedgerObject(
-                    std::move(*obj.mutable_key()), request_.ledger().sequence(), std::move(*obj.mutable_data()));
-            }
-        }
-        backend.cache().update(cacheUpdates, request_.ledger().sequence(), cacheOnly);
-        log_.debug() << "Wrote " << numObjects << " objects. Got more: " << (more ? "YES" : "NO");
-
-        return more ? CallStatus::MORE : CallStatus::DONE;
-    }
-
-    void
-    call(std::unique_ptr<org::xrpl::rpc::v1::XRPLedgerAPIService::Stub>& stub, grpc::CompletionQueue& cq)
-    {
-        context_ = std::make_unique<grpc::ClientContext>();
-
-        std::unique_ptr<grpc::ClientAsyncResponseReader<org::xrpl::rpc::v1::GetLedgerDataResponse>> rpc(
-            stub->PrepareAsyncGetLedgerData(context_.get(), request_, &cq));
-
-        rpc->StartCall();
-
-        rpc->Finish(next_.get(), &status_, this);
-    }
-
-    std::string
-    getMarkerPrefix()
-    {
-        if (next_->marker().size() == 0)
-            return "";
-        else
-            return ripple::strHex(std::string{next_->marker().data()[0]});
-    }
-
-    std::string
-    getLastKey()
-    {
-        return lastKey_;
-    }
-};
-
-template <class Derived>
-std::pair<std::vector<std::string>, bool>
-SourceImpl<Derived>::loadInitialLedger(uint32_t sequence, uint32_t numMarkers, bool cacheOnly)
-{
-    if (!stub_)
-        return {{}, false};
-
-    grpc::CompletionQueue cq;
-    void* tag;
-    bool ok = false;
-    std::vector<AsyncCallData> calls;
-    auto markers = getMarkers(numMarkers);
-
-    for (size_t i = 0; i < markers.size(); ++i)
-    {
-        std::optional<ripple::uint256> nextMarker;
-
-        if (i + 1 < markers.size())
-            nextMarker = markers[i + 1];
-
-        calls.emplace_back(sequence, markers[i], nextMarker);
-    }
-
-    log_.debug() << "Starting data download for ledger " << sequence << ". Using source = " << toString();
-
-    for (auto& c : calls)
-        c.call(stub_, cq);
-
-    size_t numFinished = 0;
-    bool abort = false;
-    size_t incr = 500000;
-    size_t progress = incr;
-    std::vector<std::string> edgeKeys;
-
-    while (numFinished < calls.size() && cq.Next(&tag, &ok))
-    {
-        assert(tag);
-        auto ptr = static_cast<AsyncCallData*>(tag);
-
-        if (!ok)
-        {
-            log_.error() << "loadInitialLedger - ok is false";
-            return {{}, false};  // handle cancelled
-        }
-        else
-        {
-            log_.trace() << "Marker prefix = " << ptr->getMarkerPrefix();
-
-            auto result = ptr->process(stub_, cq, *backend_, abort, cacheOnly);
-            if (result != AsyncCallData::CallStatus::MORE)
-            {
-                numFinished++;
-                log_.debug() << "Finished a marker. "
-                             << "Current number of finished = " << numFinished;
-
-                std::string lastKey = ptr->getLastKey();
-
-                if (lastKey.size())
-                    edgeKeys.push_back(ptr->getLastKey());
-            }
-
-            if (result == AsyncCallData::CallStatus::ERRORED)
-                abort = true;
-
-            if (backend_->cache().size() > progress)
-            {
-                log_.info() << "Downloaded " << backend_->cache().size() << " records from rippled";
-                progress += incr;
-            }
-        }
-    }
-
-    log_.info() << "Finished loadInitialLedger. cache size = " << backend_->cache().size();
-    return {std::move(edgeKeys), !abort};
-}
-
-template <class Derived>
-std::pair<grpc::Status, org::xrpl::rpc::v1::GetLedgerResponse>
-SourceImpl<Derived>::fetchLedger(uint32_t ledgerSequence, bool getObjects, bool getObjectNeighbors)
-{
-    org::xrpl::rpc::v1::GetLedgerResponse response;
-    if (!stub_)
-        return {{grpc::StatusCode::INTERNAL, "No Stub"}, response};
-
-    // ledger header with txns and metadata
-    org::xrpl::rpc::v1::GetLedgerRequest request;
-    grpc::ClientContext context;
-    request.mutable_ledger()->set_sequence(ledgerSequence);
-    request.set_transactions(true);
-    request.set_expand(true);
-    request.set_get_objects(getObjects);
-    request.set_get_object_neighbors(getObjectNeighbors);
-    request.set_user("ETL");
-    grpc::Status status = stub_->GetLedger(&context, request, &response);
-    if (status.ok() && !response.is_unlimited())
-    {
-        log_.warn() << "SourceImpl::fetchLedger - is_unlimited is "
-                       "false. Make sure secure_gateway is set "
-                       "correctly on the ETL source. source = "
-                    << toString() << " status = " << status.error_message();
-    }
-    return {status, std::move(response)};
-}
-
-template <class Derived>
-std::optional<boost::json::object>
-SourceImpl<Derived>::forwardToRippled(
-    boost::json::object const& request,
-    std::string const& clientIp,
-    boost::asio::yield_context& yield) const
-{
-    if (auto resp = forwardCache_.get(request); resp)
-    {
-        log_.debug() << "request hit forwardCache";
-        return resp;
-    }
-
-    return requestFromRippled(request, clientIp, yield);
-}
-
-template <class Derived>
-std::optional<boost::json::object>
-SourceImpl<Derived>::requestFromRippled(
-    boost::json::object const& request,
-    std::string const& clientIp,
-    boost::asio::yield_context& yield) const
-{
-    log_.trace() << "Attempting to forward request to tx. "
-                 << "request = " << boost::json::serialize(request);
-
-    boost::json::object response;
-    if (!connected_)
-    {
-        log_.error() << "Attempted to proxy but failed to connect to tx";
-        return {};
-    }
-    namespace beast = boost::beast;          // from <boost/beast.hpp>
-    namespace http = beast::http;            // from <boost/beast/http.hpp>
-    namespace websocket = beast::websocket;  // from
-    namespace net = boost::asio;             // from
-    using tcp = boost::asio::ip::tcp;        // from
-    try
-    {
-        boost::beast::error_code ec;
-        // These objects perform our I/O
-        tcp::resolver resolver{ioc_};
-
-        log_.trace() << "Creating websocket";
-        auto ws = std::make_unique<websocket::stream<beast::tcp_stream>>(ioc_);
-
-        // Look up the domain name
-        auto const results = resolver.async_resolve(ip_, wsPort_, yield[ec]);
-        if (ec)
-            return {};
-
-        ws->next_layer().expires_after(std::chrono::seconds(3));
-
-        log_.trace() << "Connecting websocket";
-        // Make the connection on the IP address we get from a lookup
-        ws->next_layer().async_connect(results, yield[ec]);
-        if (ec)
-            return {};
-
-        // Set a decorator to change the User-Agent of the handshake
-        // and to tell rippled to charge the client IP for RPC
-        // resources. See "secure_gateway" in
-        //
-        // https://github.com/ripple/rippled/blob/develop/cfg/rippled-example.cfg
-        ws->set_option(websocket::stream_base::decorator([&clientIp](websocket::request_type& req) {
-            req.set(http::field::user_agent, std::string(BOOST_BEAST_VERSION_STRING) + " websocket-client-coro");
-            req.set(http::field::forwarded, "for=" + clientIp);
-        }));
-        log_.trace() << "client ip: " << clientIp;
-
-        log_.trace() << "Performing websocket handshake";
-        // Perform the websocket handshake
-        ws->async_handshake(ip_, "/", yield[ec]);
-        if (ec)
-            return {};
-
-        log_.trace() << "Sending request";
-        // Send the message
-        ws->async_write(net::buffer(boost::json::serialize(request)), yield[ec]);
-        if (ec)
-            return {};
-
-        beast::flat_buffer buffer;
-        ws->async_read(buffer, yield[ec]);
-        if (ec)
-            return {};
-
-        auto begin = static_cast<char const*>(buffer.data().data());
-        auto end = begin + buffer.data().size();
-        auto parsed = boost::json::parse(std::string(begin, end));
-
-        if (!parsed.is_object())
-        {
-            log_.error() << "Error parsing response: " << std::string{begin, end};
-            return {};
-        }
-        log_.trace() << "Successfully forward request";
-
-        response = parsed.as_object();
-
-        response["forwarded"] = true;
-        return response;
-    }
-    catch (std::exception const& e)
-    {
-        log_.error() << "Encountered exception : " << e.what();
-        return {};
-    }
-}
+}  // namespace etl

src/etl/SystemState.h

@@ -21,30 +21,33 @@
 
 #include <atomic>
 
+namespace etl {
+
+/**
+ * @brief Represents the state of the ETL subsystem.
+ */
 struct SystemState
 {
     /**
-     * @brief Whether the process is in strict read-only mode
+     * @brief Whether the process is in strict read-only mode.
      *
      * In strict read-only mode, the process will never attempt to become the ETL writer, and will only publish ledgers
      * as they are written to the database.
      */
     bool isReadOnly = false;
 
+    std::atomic_bool isWriting = false;     /**< @brief Whether the process is writing to the database. */
+    std::atomic_bool isStopping = false;    /**< @brief Whether the software is stopping. */
+    std::atomic_bool writeConflict = false; /**< @brief Whether a write conflict was detected. */
+
     /**
-     * @brief Whether the process is writing to the database.
+     * @brief Whether clio detected an amendment block.
      *
-     * Used by server_info
+     * Being amendment blocked means that Clio was compiled with libxrpl that does not yet support some field that
+     * arrived from rippled and therefore can't extract the ledger diff. When this happens, Clio can't proceed with ETL
+     * and should log this error and only handle RPC requests.
      */
-    std::atomic_bool isWriting = false;
-
-    /**
-     * @brief Whether the software is stopping
-     */
-    std::atomic_bool isStopping = false;
-
-    /**
-     * @brief Whether a write conflict was detected
-     */
-    std::atomic_bool writeConflict = false;
+    std::atomic_bool isAmendmentBlocked = false;
 };
+
+}  // namespace etl

src/etl/impl/AmendmentBlock.h

@@ -0,0 +1,96 @@
+//------------------------------------------------------------------------------
+/*
+    This file is part of clio: https://github.com/XRPLF/clio
+    Copyright (c) 2023, the clio developers.
+
+    Permission to use, copy, modify, and distribute this software for any
+    purpose with or without fee is hereby granted, provided that the above
+    copyright notice and this permission notice appear in all copies.
+
+    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
+    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
+    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+//==============================================================================
+
+#pragma once
+
+#include <etl/SystemState.h>
+#include <util/log/Logger.h>
+
+#include <boost/asio/io_context.hpp>
+#include <boost/asio/steady_timer.hpp>
+
+#include <chrono>
+#include <functional>
+
+namespace etl::detail {
+
+struct AmendmentBlockAction
+{
+    void
+    operator()()
+    {
+        static util::Logger log{"ETL"};
+        LOG(log.fatal())
+            << "Can't process new ledgers: The current ETL source is not compatible with the version of the "
+               "libxrpl Clio is currently using. Please upgrade Clio to a newer version.";
+    }
+};
+
+template <typename ActionCallableType = AmendmentBlockAction>
+class AmendmentBlockHandler
+{
+    std::reference_wrapper<boost::asio::io_context> ctx_;
+    std::reference_wrapper<SystemState> state_;
+    boost::asio::steady_timer timer_;
+    std::chrono::milliseconds interval_;
+
+    ActionCallableType action_;
+
+public:
+    template <typename DurationType = std::chrono::seconds>
+    AmendmentBlockHandler(
+        boost::asio::io_context& ioc,
+        SystemState& state,
+        DurationType interval = DurationType{1},
+        ActionCallableType&& action = ActionCallableType())
+        : ctx_{std::ref(ioc)}
+        , state_{std::ref(state)}
+        , timer_{ioc}
+        , interval_{std::chrono::duration_cast<std::chrono::milliseconds>(interval)}
+        , action_{std::move(action)}
+    {
+    }
+
+    ~AmendmentBlockHandler()
+    {
+        boost::asio::post(ctx_.get(), [this]() { timer_.cancel(); });
+    }
+
+    void
+    onAmendmentBlock()
+    {
+        state_.get().isAmendmentBlocked = true;
+        startReportingTimer();
+    }
+
+private:
+    void
+    startReportingTimer()
+    {
+        action_();
+
+        timer_.expires_after(interval_);
+        timer_.async_wait([this](auto ec) {
+            if (!ec)
+                boost::asio::post(ctx_.get(), [this] { startReportingTimer(); });
+        });
+    }
+};
+
+}  // namespace etl::detail

src/etl/impl/AsyncData.h

@@ -0,0 +1,182 @@
+//------------------------------------------------------------------------------
+/*
+    This file is part of clio: https://github.com/XRPLF/clio
+    Copyright (c) 2023, the clio developers.
+
+    Permission to use, copy, modify, and distribute this software for any
+    purpose with or without fee is hereby granted, provided that the above
+    copyright notice and this permission notice appear in all copies.
+
+    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
+    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
+    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+//==============================================================================
+
+#pragma once
+
+#include <etl/NFTHelpers.h>
+#include <util/log/Logger.h>
+
+#include <ripple/proto/org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h>
+#include <grpcpp/grpcpp.h>
+
+namespace etl::detail {
+
+class AsyncCallData
+{
+    util::Logger log_{"ETL"};
+
+    std::unique_ptr<org::xrpl::rpc::v1::GetLedgerDataResponse> cur_;
+    std::unique_ptr<org::xrpl::rpc::v1::GetLedgerDataResponse> next_;
+
+    org::xrpl::rpc::v1::GetLedgerDataRequest request_;
+    std::unique_ptr<grpc::ClientContext> context_;
+
+    grpc::Status status_;
+    unsigned char nextPrefix_;
+
+    std::string lastKey_;
+
+public:
+    AsyncCallData(uint32_t seq, ripple::uint256 const& marker, std::optional<ripple::uint256> const& nextMarker)
+    {
+        request_.mutable_ledger()->set_sequence(seq);
+        if (marker.isNonZero())
+        {
+            request_.set_marker(marker.data(), marker.size());
+        }
+        request_.set_user("ETL");
+        nextPrefix_ = 0x00;
+        if (nextMarker)
+            nextPrefix_ = nextMarker->data()[0];
+
+        unsigned char prefix = marker.data()[0];
+
+        LOG(log_.debug()) << "Setting up AsyncCallData. marker = " << ripple::strHex(marker)
+                          << " . prefix = " << ripple::strHex(std::string(1, prefix))
+                          << " . nextPrefix_ = " << ripple::strHex(std::string(1, nextPrefix_));
+
+        assert(nextPrefix_ > prefix || nextPrefix_ == 0x00);
+
+        cur_ = std::make_unique<org::xrpl::rpc::v1::GetLedgerDataResponse>();
+        next_ = std::make_unique<org::xrpl::rpc::v1::GetLedgerDataResponse>();
+        context_ = std::make_unique<grpc::ClientContext>();
+    }
+
+    enum class CallStatus { MORE, DONE, ERRORED };
+
+    CallStatus
+    process(
+        std::unique_ptr<org::xrpl::rpc::v1::XRPLedgerAPIService::Stub>& stub,
+        grpc::CompletionQueue& cq,
+        BackendInterface& backend,
+        bool abort,
+        bool cacheOnly = false)
+    {
+        LOG(log_.trace()) << "Processing response. "
+                          << "Marker prefix = " << getMarkerPrefix();
+        if (abort)
+        {
+            LOG(log_.error()) << "AsyncCallData aborted";
+            return CallStatus::ERRORED;
+        }
+        if (!status_.ok())
+        {
+            LOG(log_.error()) << "AsyncCallData status_ not ok: "
+                              << " code = " << status_.error_code() << " message = " << status_.error_message();
+            return CallStatus::ERRORED;
+        }
+        if (!next_->is_unlimited())
+        {
+            LOG(log_.warn()) << "AsyncCallData is_unlimited is false. Make sure "
+                                "secure_gateway is set correctly at the ETL source";
+        }
+
+        std::swap(cur_, next_);
+
+        bool more = true;
+
+        // if no marker returned, we are done
+        if (cur_->marker().size() == 0)
+            more = false;
+
+        // if returned marker is greater than our end, we are done
+        unsigned char prefix = cur_->marker()[0];
+        if (nextPrefix_ != 0x00 && prefix >= nextPrefix_)
+            more = false;
+
+        // if we are not done, make the next async call
+        if (more)
+        {
+            request_.set_marker(std::move(cur_->marker()));
+            call(stub, cq);
+        }
+
+        auto const numObjects = cur_->ledger_objects().objects_size();
+        LOG(log_.debug()) << "Writing " << numObjects << " objects";
+
+        std::vector<data::LedgerObject> cacheUpdates;
+        cacheUpdates.reserve(numObjects);
+
+        for (int i = 0; i < numObjects; ++i)
+        {
+            auto& obj = *(cur_->mutable_ledger_objects()->mutable_objects(i));
+            if (!more && nextPrefix_ != 0x00)
+            {
+                if (static_cast<unsigned char>(obj.key()[0]) >= nextPrefix_)
+                    continue;
+            }
+            cacheUpdates.push_back(
+                {*ripple::uint256::fromVoidChecked(obj.key()),
+                 {obj.mutable_data()->begin(), obj.mutable_data()->end()}});
+            if (!cacheOnly)
+            {
+                if (lastKey_.size())
+                    backend.writeSuccessor(std::move(lastKey_), request_.ledger().sequence(), std::string{obj.key()});
+                lastKey_ = obj.key();
+                backend.writeNFTs(getNFTDataFromObj(request_.ledger().sequence(), obj.key(), obj.data()));
+                backend.writeLedgerObject(
+                    std::move(*obj.mutable_key()), request_.ledger().sequence(), std::move(*obj.mutable_data()));
+            }
+        }
+        backend.cache().update(cacheUpdates, request_.ledger().sequence(), cacheOnly);
+        LOG(log_.debug()) << "Wrote " << numObjects << " objects. Got more: " << (more ? "YES" : "NO");
+
+        return more ? CallStatus::MORE : CallStatus::DONE;
+    }
+
+    void
+    call(std::unique_ptr<org::xrpl::rpc::v1::XRPLedgerAPIService::Stub>& stub, grpc::CompletionQueue& cq)
+    {
+        context_ = std::make_unique<grpc::ClientContext>();
+
+        std::unique_ptr<grpc::ClientAsyncResponseReader<org::xrpl::rpc::v1::GetLedgerDataResponse>> rpc(
+            stub->PrepareAsyncGetLedgerData(context_.get(), request_, &cq));
+
+        rpc->StartCall();
+
+        rpc->Finish(next_.get(), &status_, this);
+    }
+
+    std::string
+    getMarkerPrefix()
+    {
+        if (next_->marker().size() == 0)
+            return "";
+        else
+            return ripple::strHex(std::string{next_->marker().data()[0]});
+    }
+
+    std::string
+    getLastKey()
+    {
+        return lastKey_;
+    }
+};
+
+}  // namespace etl::detail

src/etl/impl/CacheLoader.h

@@ -19,22 +19,22 @@
 
 #pragma once
 
-#include <log/Logger.h>
+#include <data/BackendInterface.h>
+#include <util/log/Logger.h>
 
-#include <ripple/ledger/ReadView.h>
+#include <ripple/proto/org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h>
 #include <boost/algorithm/string.hpp>
 #include <boost/asio/spawn.hpp>
 #include <boost/beast/core.hpp>
 #include <boost/beast/core/string.hpp>
 #include <boost/beast/websocket.hpp>
-#include "org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h"
 #include <grpcpp/grpcpp.h>
 
 #include <chrono>
 #include <mutex>
 #include <thread>
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief Cache loading interface
@@ -44,7 +44,7 @@ class CacheLoader
 {
     enum class LoadStyle { ASYNC, SYNC, NOT_AT_ALL };
 
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::reference_wrapper<boost::asio::io_context> ioContext_;
     std::shared_ptr<BackendInterface> backend_;
@@ -73,7 +73,7 @@ class CacheLoader
 
 public:
     CacheLoader(
-        clio::Config const& config,
+        util::Config const& config,
         boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> const& backend,
         CacheType& ledgerCache)
@@ -132,7 +132,7 @@ public:
         if (cacheLoadStyle_ == LoadStyle::NOT_AT_ALL)
         {
             cache_.get().setDisabled();
-            log_.warn() << "Cache is disabled. Not loading";
+            LOG(log_.warn()) << "Cache is disabled. Not loading";
             return;
         }
 
@@ -165,10 +165,10 @@ public:
         // If loading synchronously, poll cache until full
         while (cacheLoadStyle_ == LoadStyle::SYNC && not cache_.get().isFull())
         {
-            log_.debug() << "Cache not full. Cache size = " << cache_.get().size() << ". Sleeping ...";
+            LOG(log_.debug()) << "Cache not full. Cache size = " << cache_.get().size() << ". Sleeping ...";
             std::this_thread::sleep_for(std::chrono::seconds(10));
             if (cache_.get().isFull())
-                log_.info() << "Cache is full. Cache size = " << cache_.get().size();
+                LOG(log_.info()) << "Cache is full. Cache size = " << cache_.get().size();
         }
     }
 
@@ -184,9 +184,9 @@ private:
         uint32_t ledgerIndex,
         std::string const& ip,
         std::string const& port,
-        boost::asio::yield_context& yield)
+        boost::asio::yield_context yield)
     {
-        log_.info() << "Loading cache from peer. ip = " << ip << " . port = " << port;
+        LOG(log_.info()) << "Loading cache from peer. ip = " << ip << " . port = " << port;
         namespace beast = boost::beast;          // from <boost/beast.hpp>
         namespace http = beast::http;            // from <boost/beast/http.hpp>
         namespace websocket = beast::websocket;  // from
@@ -198,7 +198,7 @@ private:
             // These objects perform our I/O
             tcp::resolver resolver{ioContext_.get()};
 
-            log_.trace() << "Creating websocket";
+            LOG(log_.trace()) << "Creating websocket";
             auto ws = std::make_unique<websocket::stream<beast::tcp_stream>>(ioContext_.get());
 
             // Look up the domain name
@@ -206,13 +206,13 @@ private:
             if (ec)
                 return {};
 
-            log_.trace() << "Connecting websocket";
+            LOG(log_.trace()) << "Connecting websocket";
             // Make the connection on the IP address we get from a lookup
             ws->next_layer().async_connect(results, yield[ec]);
             if (ec)
                 return false;
 
-            log_.trace() << "Performing websocket handshake";
+            LOG(log_.trace()) << "Performing websocket handshake";
             // Perform the websocket handshake
             ws->async_handshake(ip, "/", yield[ec]);
             if (ec)
@@ -220,7 +220,7 @@ private:
 
             std::optional<boost::json::value> marker;
 
-            log_.trace() << "Sending request";
+            LOG(log_.trace()) << "Sending request";
             auto getRequest = [&](auto marker) {
                 boost::json::object request = {
                     {"command", "ledger_data"},
@@ -242,7 +242,7 @@ private:
                 ws->async_write(net::buffer(boost::json::serialize(getRequest(marker))), yield[ec]);
                 if (ec)
                 {
-                    log_.error() << "error writing = " << ec.message();
+                    LOG(log_.error()) << "error writing = " << ec.message();
                     return false;
                 }
 
@@ -250,7 +250,7 @@ private:
                 ws->async_read(buffer, yield[ec]);
                 if (ec)
                 {
-                    log_.error() << "error reading = " << ec.message();
+                    LOG(log_.error()) << "error reading = " << ec.message();
                     return false;
                 }
 
@@ -259,27 +259,28 @@ private:
 
                 if (!parsed.is_object())
                 {
-                    log_.error() << "Error parsing response: " << raw;
+                    LOG(log_.error()) << "Error parsing response: " << raw;
                     return false;
                 }
-                log_.trace() << "Successfully parsed response " << parsed;
+                LOG(log_.trace()) << "Successfully parsed response " << parsed;
 
                 if (auto const& response = parsed.as_object(); response.contains("error"))
                 {
-                    log_.error() << "Response contains error: " << response;
+                    LOG(log_.error()) << "Response contains error: " << response;
                     auto const& err = response.at("error");
                     if (err.is_string() && err.as_string() == "lgrNotFound")
                     {
                         ++numAttempts;
                         if (numAttempts >= 5)
                         {
-                            log_.error() << " ledger not found at peer after 5 attempts. "
-                                            "peer = "
-                                         << ip << " ledger = " << ledgerIndex
-                                         << ". Check your config and the health of the peer";
+                            LOG(log_.error()) << " ledger not found at peer after 5 attempts. "
+                                                 "peer = "
+                                              << ip << " ledger = " << ledgerIndex
+                                              << ". Check your config and the health of the peer";
                             return false;
                         }
-                        log_.warn() << "Ledger not found. ledger = " << ledgerIndex << ". Sleeping and trying again";
+                        LOG(log_.warn()) << "Ledger not found. ledger = " << ledgerIndex
+                                         << ". Sleeping and trying again";
                         std::this_thread::sleep_for(std::chrono::seconds(1));
                         continue;
                     }
@@ -290,7 +291,7 @@ private:
 
                 if (!response.contains("cache_full") || !response.at("cache_full").as_bool())
                 {
-                    log_.error() << "cache not full for clio node. ip = " << ip;
+                    LOG(log_.error()) << "cache not full for clio node. ip = " << ip;
                     return false;
                 }
                 if (response.contains("marker"))
@@ -300,17 +301,17 @@ private:
 
                 auto const& state = response.at("state").as_array();
 
-                std::vector<Backend::LedgerObject> objects;
+                std::vector<data::LedgerObject> objects;
                 objects.reserve(state.size());
                 for (auto const& ledgerObject : state)
                 {
                     auto const& obj = ledgerObject.as_object();
 
-                    Backend::LedgerObject stateObject = {};
+                    data::LedgerObject stateObject = {};
 
                     if (!stateObject.key.parseHex(obj.at("index").as_string().c_str()))
                     {
-                        log_.error() << "failed to parse object id";
+                        LOG(log_.error()) << "failed to parse object id";
                         return false;
                     }
                     boost::algorithm::unhex(obj.at("data").as_string().c_str(), std::back_inserter(stateObject.blob));
@@ -319,17 +320,17 @@ private:
                 cache_.get().update(objects, ledgerIndex, true);
 
                 if (marker)
-                    log_.debug() << "At marker " << *marker;
+                    LOG(log_.debug()) << "At marker " << *marker;
             } while (marker || !started);
 
-            log_.info() << "Finished downloading ledger from clio node. ip = " << ip;
+            LOG(log_.info()) << "Finished downloading ledger from clio node. ip = " << ip;
 
             cache_.get().setFull();
             return true;
         }
         catch (std::exception const& e)
         {
-            log_.error() << "Encountered exception : " << e.what() << " - ip = " << ip;
+            LOG(log_.error()) << "Encountered exception : " << e.what() << " - ip = " << ip;
             return false;
         }
     }
@@ -337,14 +338,14 @@ private:
     void
     loadCacheFromDb(uint32_t seq)
     {
-        std::vector<Backend::LedgerObject> diff;
+        std::vector<data::LedgerObject> diff;
         std::vector<std::optional<ripple::uint256>> cursors;
 
         auto append = [](auto&& a, auto&& b) { a.insert(std::end(a), std::begin(b), std::end(b)); };
 
         for (size_t i = 0; i < numCacheDiffs_; ++i)
         {
-            append(diff, Backend::synchronousAndRetryOnTimeout([&](auto yield) {
+            append(diff, data::synchronousAndRetryOnTimeout([&](auto yield) {
                        return backend_->fetchLedgerDiff(seq - i, yield);
                    }));
         }
@@ -356,28 +357,28 @@ private:
         diff.erase(std::unique(diff.begin(), diff.end(), [](auto a, auto b) { return a.key == b.key; }), diff.end());
 
         cursors.push_back({});
-        for (auto& obj : diff)
+        for (auto const& obj : diff)
             if (obj.blob.size())
                 cursors.push_back({obj.key});
         cursors.push_back({});
 
         std::stringstream cursorStr;
-        for (auto& c : cursors)
+        for (auto const& c : cursors)
             if (c)
                 cursorStr << ripple::strHex(*c) << ", ";
 
-        log_.info() << "Loading cache. num cursors = " << cursors.size() - 1;
-        log_.trace() << "cursors = " << cursorStr.str();
+        LOG(log_.info()) << "Loading cache. num cursors = " << cursors.size() - 1;
+        LOG(log_.trace()) << "cursors = " << cursorStr.str();
 
-        thread_ = std::thread{[this, seq, cursors]() {
+        thread_ = std::thread{[this, seq, cursors = std::move(cursors)]() {
             auto startTime = std::chrono::system_clock::now();
             auto markers = std::make_shared<std::atomic_int>(0);
             auto numRemaining = std::make_shared<std::atomic_int>(cursors.size() - 1);
 
             for (size_t i = 0; i < cursors.size() - 1; ++i)
             {
-                auto const start = cursors[i];
-                auto const end = cursors[i + 1];
+                auto const start = cursors.at(i);
+                auto const end = cursors.at(i + 1);
 
                 markers->wait(numCacheMarkers_);
                 ++(*markers);
@@ -385,14 +386,14 @@ private:
                 boost::asio::spawn(
                     ioContext_.get(),
                     [this, seq, start, end, numRemaining, startTime, markers](boost::asio::yield_context yield) {
-                        std::optional<ripple::uint256> cursor = start;
+                        auto cursor = start;
                         std::string cursorStr =
-                            cursor.has_value() ? ripple::strHex(cursor.value()) : ripple::strHex(Backend::firstKey);
-                        log_.debug() << "Starting a cursor: " << cursorStr << " markers = " << *markers;
+                            cursor.has_value() ? ripple::strHex(cursor.value()) : ripple::strHex(data::firstKey);
+                        LOG(log_.debug()) << "Starting a cursor: " << cursorStr << " markers = " << *markers;
 
                         while (not stopping_)
                         {
-                            auto res = Backend::retryOnTimeout([this, seq, &cursor, &yield]() {
+                            auto res = data::retryOnTimeout([this, seq, &cursor, yield]() {
                                 return backend_->fetchLedgerPage(cursor, seq, cachePageFetchSize_, false, yield);
                             });
 
@@ -401,9 +402,9 @@ private:
                             if (!res.cursor || (end && *(res.cursor) > *end))
                                 break;
 
-                            log_.trace() << "Loading cache. cache size = " << cache_.get().size()
-                                         << " - cursor = " << ripple::strHex(res.cursor.value())
-                                         << " start = " << cursorStr << " markers = " << *markers;
+                            LOG(log_.trace()) << "Loading cache. cache size = " << cache_.get().size()
+                                              << " - cursor = " << ripple::strHex(res.cursor.value())
+                                              << " start = " << cursorStr << " markers = " << *markers;
 
                             cursor = std::move(res.cursor);
                         }
@@ -416,14 +417,14 @@ private:
                             auto endTime = std::chrono::system_clock::now();
                             auto duration = std::chrono::duration_cast<std::chrono::seconds>(endTime - startTime);
 
-                            log_.info() << "Finished loading cache. cache size = " << cache_.get().size() << ". Took "
-                                        << duration.count() << " seconds";
+                            LOG(log_.info()) << "Finished loading cache. cache size = " << cache_.get().size()
+                                             << ". Took " << duration.count() << " seconds";
                             cache_.get().setFull();
                         }
                         else
                         {
-                            log_.info() << "Finished a cursor. num remaining = " << *numRemaining
-                                        << " start = " << cursorStr << " markers = " << *markers;
+                            LOG(log_.info()) << "Finished a cursor. num remaining = " << *numRemaining
+                                             << " start = " << cursorStr << " markers = " << *markers;
                         }
                     });
             }
@@ -431,4 +432,4 @@ private:
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/ExtractionDataPipe.h

@@ -20,12 +20,12 @@
 #pragma once
 
 #include <etl/ETLHelpers.h>
-#include <log/Logger.h>
+#include <util/log/Logger.h>
 
 #include <memory>
 #include <vector>
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief A collection of thread safe async queues used by Extractor and Transformer to communicate
@@ -40,7 +40,7 @@ public:
     constexpr static auto TOTAL_MAX_IN_QUEUE = 1000u;
 
 private:
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     uint32_t stride_;
     uint32_t startSequence_;
@@ -126,9 +126,9 @@ private:
     std::shared_ptr<QueueType>
     getQueue(uint32_t sequence)
     {
-        log_.debug() << "Grabbing extraction queue for " << sequence << "; start was " << startSequence_;
+        LOG(log_.debug()) << "Grabbing extraction queue for " << sequence << "; start was " << startSequence_;
         return queues_[(sequence - startSequence_) % stride_];
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/Extractor.h

@@ -20,8 +20,8 @@
 #pragma once
 
 #include <etl/SystemState.h>
-#include <log/Logger.h>
 #include <util/Profiler.h>
+#include <util/log/Logger.h>
 
 #include <ripple/beast/core/CurrentThreadName.h>
 
@@ -29,7 +29,7 @@
 #include <mutex>
 #include <thread>
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief Extractor thread that is fetching GRPC data and enqueue it on the DataPipeType
@@ -37,7 +37,7 @@ namespace clio::detail {
 template <typename DataPipeType, typename NetworkValidatedLedgersType, typename LedgerFetcherType>
 class Extractor
 {
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::reference_wrapper<DataPipeType> pipe_;
     std::shared_ptr<NetworkValidatedLedgersType> networkValidatedLedgers_;
@@ -90,7 +90,7 @@ private:
 
         while (!shouldFinish(currentSequence) && networkValidatedLedgers_->waitUntilValidatedByNetwork(currentSequence))
         {
-            auto [fetchResponse, time] = util::timed<std::chrono::duration<double>>(
+            auto [fetchResponse, time] = ::util::timed<std::chrono::duration<double>>(
                 [this, currentSequence]() { return ledgerFetcher_.get().fetchDataAndDiff(currentSequence); });
             totalTime += time;
 
@@ -103,9 +103,9 @@ private:
 
             // TODO: extract this part into a strategy perhaps
             auto const tps = fetchResponse->transactions_list().transactions_size() / time;
-            log_.info() << "Extract phase time = " << time << "; Extract phase tps = " << tps
-                        << "; Avg extract time = " << totalTime / (currentSequence - startSequence_ + 1)
-                        << "; seq = " << currentSequence;
+            LOG(log_.info()) << "Extract phase time = " << time << "; Extract phase tps = " << tps
+                             << "; Avg extract time = " << totalTime / (currentSequence - startSequence_ + 1)
+                             << "; seq = " << currentSequence;
 
             pipe_.get().push(currentSequence, std::move(fetchResponse));
             currentSequence += pipe_.get().getStride();
@@ -137,4 +137,4 @@ private:
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/ForwardCache.cpp

@@ -24,12 +24,12 @@
 #include <boost/asio/spawn.hpp>
 #include <boost/json.hpp>
 
-namespace clio::detail {
+namespace etl::detail {
 
 void
 ForwardCache::freshen()
 {
-    log_.trace() << "Freshening ForwardCache";
+    LOG(log_.trace()) << "Freshening ForwardCache";
 
     auto numOutstanding = std::make_shared<std::atomic_uint>(latestForwarded_.size());
 
@@ -70,7 +70,7 @@ ForwardCache::get(boost::json::object const& request) const
 
     if (!command)
         return {};
-    if (RPC::specifiesCurrentOrClosedLedger(request))
+    if (rpc::specifiesCurrentOrClosedLedger(request))
         return {};
 
     std::shared_lock lk(mtx_);
@@ -80,4 +80,4 @@ ForwardCache::get(boost::json::object const& request) const
     return {latestForwarded_.at(*command)};
 }
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/ForwardCache.h

@@ -19,10 +19,10 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
-#include <config/Config.h>
+#include <data/BackendInterface.h>
 #include <etl/ETLHelpers.h>
-#include <log/Logger.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
 
 #include <boost/asio.hpp>
 #include <boost/json.hpp>
@@ -33,7 +33,7 @@
 
 class Source;
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief Cache for rippled responses
@@ -42,20 +42,20 @@ class ForwardCache
 {
     using ResponseType = std::optional<boost::json::object>;
 
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     mutable std::shared_mutex mtx_;
     std::unordered_map<std::string, ResponseType> latestForwarded_;
-    boost::asio::io_context::strand strand_;
-    Source const& source_;
+    boost::asio::strand<boost::asio::io_context::executor_type> strand_;
+    etl::Source const& source_;
     std::uint32_t duration_ = 10;
 
     void
     clear();
 
 public:
-    ForwardCache(clio::Config const& config, boost::asio::io_context& ioc, Source const& source)
-        : strand_(ioc), source_(source)
+    ForwardCache(util::Config const& config, boost::asio::io_context& ioc, Source const& source)
+        : strand_(boost::asio::make_strand(ioc)), source_(source)
     {
         if (config.contains("cache"))
         {
@@ -79,4 +79,4 @@ public:
     get(boost::json::object const& command) const;
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/LedgerFetcher.h

@@ -19,17 +19,16 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
+#include <data/BackendInterface.h>
 #include <etl/Source.h>
-#include <log/Logger.h>
+#include <util/log/Logger.h>
 
-#include <ripple/ledger/ReadView.h>
-#include "org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h"
+#include <ripple/proto/org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h>
 #include <grpcpp/grpcpp.h>
 
 #include <optional>
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief GRPC Ledger data fetcher
@@ -41,7 +40,7 @@ public:
     using OptionalGetLedgerResponseType = typename LoadBalancerType::OptionalGetLedgerResponseType;
 
 private:
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::shared_ptr<BackendInterface> backend_;
     std::shared_ptr<LoadBalancerType> loadBalancer_;
@@ -65,13 +64,13 @@ public:
      * @return ledger header and transaction+metadata blobs; empty optional if the server is shutting down
      */
     OptionalGetLedgerResponseType
-    fetchData(uint32_t seq)
+    fetchData(uint32_t sequence)
     {
-        log_.debug() << "Attempting to fetch ledger with sequence = " << seq;
+        LOG(log_.debug()) << "Attempting to fetch ledger with sequence = " << sequence;
 
-        auto response = loadBalancer_->fetchLedger(seq, false, false);
+        auto response = loadBalancer_->fetchLedger(sequence, false, false);
         if (response)
-            log_.trace() << "GetLedger reply = " << response->DebugString();
+            LOG(log_.trace()) << "GetLedger reply = " << response->DebugString();
         return response;
     }
 
@@ -86,17 +85,17 @@ public:
      * this ledger and the parent; Empty optional if the server is shutting down
      */
     OptionalGetLedgerResponseType
-    fetchDataAndDiff(uint32_t seq)
+    fetchDataAndDiff(uint32_t sequence)
     {
-        log_.debug() << "Attempting to fetch ledger with sequence = " << seq;
+        LOG(log_.debug()) << "Attempting to fetch ledger with sequence = " << sequence;
 
         auto response = loadBalancer_->fetchLedger(
-            seq, true, !backend_->cache().isFull() || backend_->cache().latestLedgerSequence() >= seq);
+            sequence, true, !backend_->cache().isFull() || backend_->cache().latestLedgerSequence() >= sequence);
         if (response)
-            log_.trace() << "GetLedger reply = " << response->DebugString();
+            LOG(log_.trace()) << "GetLedger reply = " << response->DebugString();
 
         return response;
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/LedgerLoader.h

@@ -19,19 +19,21 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
+#include <data/BackendInterface.h>
 #include <etl/NFTHelpers.h>
 #include <etl/SystemState.h>
 #include <etl/impl/LedgerFetcher.h>
-#include <log/Logger.h>
 #include <util/LedgerUtils.h>
 #include <util/Profiler.h>
+#include <util/log/Logger.h>
 
 #include <ripple/beast/core/CurrentThreadName.h>
-#include <ripple/ledger/ReadView.h>
 
 #include <memory>
 
+/**
+ * @brief Account transactions, NFT transactions and NFT data bundled togeher.
+ */
 struct FormattedTransactionsData
 {
     std::vector<AccountTransactionsData> accountTxData;
@@ -39,7 +41,7 @@ struct FormattedTransactionsData
     std::vector<NFTsData> nfTokensData;
 };
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief Loads ledger data into the DB
@@ -53,7 +55,7 @@ public:
     using RawLedgerObjectType = typename LoadBalancerType::RawLedgerObjectType;
 
 private:
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::shared_ptr<BackendInterface> backend_;
     std::shared_ptr<LoadBalancerType> loadBalancer_;
@@ -85,7 +87,7 @@ public:
      * nft_token_transactions tables (mostly transaction hashes, corresponding nodestore hashes and affected accounts)
      */
     FormattedTransactionsData
-    insertTransactions(ripple::LedgerInfo const& ledger, GetLedgerResponseType& data)
+    insertTransactions(ripple::LedgerHeader const& ledger, GetLedgerResponseType& data)
     {
         FormattedTransactionsData result;
 
@@ -96,7 +98,7 @@ public:
             ripple::SerialIter it{raw->data(), raw->size()};
             ripple::STTx sttx{it};
 
-            log_.trace() << "Inserting transaction = " << sttx.getTransactionID();
+            LOG(log_.trace()) << "Inserting transaction = " << sttx.getTransactionID();
 
             ripple::TxMeta txMeta{sttx.getTransactionID(), ledger.seq, txn.metadata_blob()};
 
@@ -105,9 +107,8 @@ public:
             if (maybeNFT)
                 result.nfTokensData.push_back(*maybeNFT);
 
-            auto journal = ripple::debugLog();
-            result.accountTxData.emplace_back(txMeta, sttx.getTransactionID(), journal);
-            std::string keyStr{(const char*)sttx.getTransactionID().data(), 32};
+            result.accountTxData.emplace_back(txMeta, sttx.getTransactionID());
+            std::string keyStr{reinterpret_cast<const char*>(sttx.getTransactionID().data()), 32};
             backend_->writeTransaction(
                 std::move(keyStr),
                 ledger.seq,
@@ -140,14 +141,14 @@ public:
      * @param sequence the sequence of the ledger to download
      * @return The ledger downloaded, with a full transaction and account state map
      */
-    std::optional<ripple::LedgerInfo>
+    std::optional<ripple::LedgerHeader>
     loadInitialLedger(uint32_t sequence)
     {
         // check that database is actually empty
         auto rng = backend_->hardFetchLedgerRangeNoThrow();
         if (rng)
         {
-            log_.fatal() << "Database is not empty";
+            LOG(log_.fatal()) << "Database is not empty";
             assert(false);
             return {};
         }
@@ -158,20 +159,20 @@ public:
         if (!ledgerData)
             return {};
 
-        ripple::LedgerInfo lgrInfo = util::deserializeHeader(ripple::makeSlice(ledgerData->ledger_header()));
+        ripple::LedgerHeader lgrInfo = ::util::deserializeHeader(ripple::makeSlice(ledgerData->ledger_header()));
 
-        log_.debug() << "Deserialized ledger header. " << util::toString(lgrInfo);
+        LOG(log_.debug()) << "Deserialized ledger header. " << ::util::toString(lgrInfo);
 
-        auto timeDiff = util::timed<std::chrono::duration<double>>([this, sequence, &lgrInfo, &ledgerData]() {
+        auto timeDiff = ::util::timed<std::chrono::duration<double>>([this, sequence, &lgrInfo, &ledgerData]() {
             backend_->startWrites();
 
-            log_.debug() << "Started writes";
+            LOG(log_.debug()) << "Started writes";
 
             backend_->writeLedger(lgrInfo, std::move(*ledgerData->mutable_ledger_header()));
 
-            log_.debug() << "Wrote ledger";
+            LOG(log_.debug()) << "Wrote ledger";
             FormattedTransactionsData insertTxResult = insertTransactions(lgrInfo, *ledgerData);
-            log_.debug() << "Inserted txns";
+            LOG(log_.debug()) << "Inserted txns";
 
             // download the full account state map. This function downloads full
             // ledger data and pushes the downloaded data into the writeQueue.
@@ -185,57 +186,59 @@ public:
                 size_t numWrites = 0;
                 backend_->cache().setFull();
 
-                auto seconds = util::timed<std::chrono::seconds>([this, edgeKeys = &edgeKeys, sequence, &numWrites]() {
-                    for (auto& key : *edgeKeys)
-                    {
-                        log_.debug() << "Writing edge key = " << ripple::strHex(key);
-                        auto succ = backend_->cache().getSuccessor(*ripple::uint256::fromVoidChecked(key), sequence);
-                        if (succ)
-                            backend_->writeSuccessor(std::move(key), sequence, uint256ToString(succ->key));
-                    }
-
-                    ripple::uint256 prev = Backend::firstKey;
-                    while (auto cur = backend_->cache().getSuccessor(prev, sequence))
-                    {
-                        assert(cur);
-                        if (prev == Backend::firstKey)
-                            backend_->writeSuccessor(uint256ToString(prev), sequence, uint256ToString(cur->key));
-
-                        if (isBookDir(cur->key, cur->blob))
+                auto seconds =
+                    ::util::timed<std::chrono::seconds>([this, edgeKeys = &edgeKeys, sequence, &numWrites]() {
+                        for (auto& key : *edgeKeys)
                         {
-                            auto base = getBookBase(cur->key);
-                            // make sure the base is not an actual object
-                            if (!backend_->cache().get(cur->key, sequence))
-                            {
-                                auto succ = backend_->cache().getSuccessor(base, sequence);
-                                assert(succ);
-                                if (succ->key == cur->key)
-                                {
-                                    log_.debug() << "Writing book successor = " << ripple::strHex(base) << " - "
-                                                 << ripple::strHex(cur->key);
-
-                                    backend_->writeSuccessor(
-                                        uint256ToString(base), sequence, uint256ToString(cur->key));
-                                }
-                            }
-
-                            ++numWrites;
+                            LOG(log_.debug()) << "Writing edge key = " << ripple::strHex(key);
+                            auto succ =
+                                backend_->cache().getSuccessor(*ripple::uint256::fromVoidChecked(key), sequence);
+                            if (succ)
+                                backend_->writeSuccessor(std::move(key), sequence, uint256ToString(succ->key));
                         }
 
-                        prev = std::move(cur->key);
-                        if (numWrites % 100000 == 0 && numWrites != 0)
-                            log_.info() << "Wrote " << numWrites << " book successors";
-                    }
+                        ripple::uint256 prev = data::firstKey;
+                        while (auto cur = backend_->cache().getSuccessor(prev, sequence))
+                        {
+                            assert(cur);
+                            if (prev == data::firstKey)
+                                backend_->writeSuccessor(uint256ToString(prev), sequence, uint256ToString(cur->key));
 
-                    backend_->writeSuccessor(uint256ToString(prev), sequence, uint256ToString(Backend::lastKey));
-                    ++numWrites;
-                });
+                            if (isBookDir(cur->key, cur->blob))
+                            {
+                                auto base = getBookBase(cur->key);
+                                // make sure the base is not an actual object
+                                if (!backend_->cache().get(cur->key, sequence))
+                                {
+                                    auto succ = backend_->cache().getSuccessor(base, sequence);
+                                    assert(succ);
+                                    if (succ->key == cur->key)
+                                    {
+                                        LOG(log_.debug()) << "Writing book successor = " << ripple::strHex(base)
+                                                          << " - " << ripple::strHex(cur->key);
 
-                log_.info() << "Looping through cache and submitting all writes took " << seconds
-                            << " seconds. numWrites = " << std::to_string(numWrites);
+                                        backend_->writeSuccessor(
+                                            uint256ToString(base), sequence, uint256ToString(cur->key));
+                                    }
+                                }
+
+                                ++numWrites;
+                            }
+
+                            prev = std::move(cur->key);
+                            if (numWrites % 100000 == 0 && numWrites != 0)
+                                LOG(log_.info()) << "Wrote " << numWrites << " book successors";
+                        }
+
+                        backend_->writeSuccessor(uint256ToString(prev), sequence, uint256ToString(data::lastKey));
+                        ++numWrites;
+                    });
+
+                LOG(log_.info()) << "Looping through cache and submitting all writes took " << seconds
+                                 << " seconds. numWrites = " << std::to_string(numWrites);
             }
 
-            log_.debug() << "Loaded initial ledger";
+            LOG(log_.debug()) << "Loaded initial ledger";
 
             if (not state_.get().isStopping)
             {
@@ -247,9 +250,9 @@ public:
             backend_->finishWrites(sequence);
         });
 
-        log_.debug() << "Time to download and store ledger = " << timeDiff;
+        LOG(log_.debug()) << "Time to download and store ledger = " << timeDiff;
         return lgrInfo;
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/LedgerPublisher.h

@@ -19,17 +19,17 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
+#include <data/BackendInterface.h>
 #include <etl/SystemState.h>
-#include <log/Logger.h>
 #include <util/LedgerUtils.h>
 #include <util/Profiler.h>
+#include <util/log/Logger.h>
 
-#include <ripple/ledger/ReadView.h>
+#include <ripple/protocol/LedgerHeader.h>
 
 #include <chrono>
 
-namespace clio::detail {
+namespace etl::detail {
 
 /**
  * @brief Publishes ledgers in a synchronized fashion.
@@ -45,9 +45,9 @@ namespace clio::detail {
 template <typename SubscriptionManagerType>
 class LedgerPublisher
 {
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
-    boost::asio::io_context::strand publishStrand_;
+    boost::asio::strand<boost::asio::io_context::executor_type> publishStrand_;
 
     std::shared_ptr<BackendInterface> backend_;
     std::shared_ptr<SubscriptionManagerType> subscriptions_;
@@ -69,9 +69,12 @@ public:
     LedgerPublisher(
         boost::asio::io_context& ioc,
         std::shared_ptr<BackendInterface> backend,
-        std::shared_ptr<SubscriptionManager> subscriptions,
+        std::shared_ptr<feed::SubscriptionManager> subscriptions,
         SystemState const& state)
-        : publishStrand_{ioc}, backend_{backend}, subscriptions_{subscriptions}, state_{std::cref(state)}
+        : publishStrand_{boost::asio::make_strand(ioc)}
+        , backend_{backend}
+        , subscriptions_{subscriptions}
+        , state_{std::cref(state)}
     {
     }
 
@@ -86,7 +89,7 @@ public:
     bool
     publish(uint32_t ledgerSequence, std::optional<uint32_t> maxAttempts)
     {
-        log_.info() << "Attempting to publish ledger = " << ledgerSequence;
+        LOG(log_.info()) << "Attempting to publish ledger = " << ledgerSequence;
         size_t numAttempts = 0;
         while (not state_.get().isStopping)
         {
@@ -94,14 +97,14 @@ public:
 
             if (!range || range->maxSequence < ledgerSequence)
             {
-                log_.debug() << "Trying to publish. Could not find "
-                                "ledger with sequence = "
-                             << ledgerSequence;
+                LOG(log_.debug()) << "Trying to publish. Could not find "
+                                     "ledger with sequence = "
+                                  << ledgerSequence;
 
                 // We try maxAttempts times to publish the ledger, waiting one second in between each attempt.
                 if (maxAttempts && numAttempts >= maxAttempts)
                 {
-                    log_.debug() << "Failed to publish ledger after " << numAttempts << " attempts.";
+                    LOG(log_.debug()) << "Failed to publish ledger after " << numAttempts << " attempts.";
                     return false;
                 }
                 std::this_thread::sleep_for(std::chrono::seconds(1));
@@ -110,7 +113,7 @@ public:
             }
             else
             {
-                auto lgr = Backend::synchronousAndRetryOnTimeout(
+                auto lgr = data::synchronousAndRetryOnTimeout(
                     [&](auto yield) { return backend_->fetchLedgerBySequence(ledgerSequence, yield); });
 
                 assert(lgr);
@@ -130,16 +133,16 @@ public:
      * @param lgrInfo the ledger to publish
      */
     void
-    publish(ripple::LedgerInfo const& lgrInfo)
+    publish(ripple::LedgerHeader const& lgrInfo)
     {
         boost::asio::post(publishStrand_, [this, lgrInfo = lgrInfo]() {
-            log_.info() << "Publishing ledger " << std::to_string(lgrInfo.seq);
+            LOG(log_.info()) << "Publishing ledger " << std::to_string(lgrInfo.seq);
 
             if (!state_.get().isWriting)
             {
-                log_.info() << "Updating cache";
+                LOG(log_.info()) << "Updating cache";
 
-                std::vector<Backend::LedgerObject> diff = Backend::synchronousAndRetryOnTimeout(
+                std::vector<data::LedgerObject> diff = data::synchronousAndRetryOnTimeout(
                     [&](auto yield) { return backend_->fetchLedgerDiff(lgrInfo.seq, yield); });
 
                 backend_->cache().update(diff, lgrInfo.seq);  // todo: inject cache to update, don't use backend cache
@@ -153,10 +156,10 @@ public:
             // TODO: this probably should be a strategy
             if (age < 600)
             {
-                std::optional<ripple::Fees> fees = Backend::synchronousAndRetryOnTimeout(
+                std::optional<ripple::Fees> fees = data::synchronousAndRetryOnTimeout(
                     [&](auto yield) { return backend_->fetchFees(lgrInfo.seq, yield); });
 
-                std::vector<Backend::TransactionAndMetadata> transactions = Backend::synchronousAndRetryOnTimeout(
+                std::vector<data::TransactionAndMetadata> transactions = data::synchronousAndRetryOnTimeout(
                     [&](auto yield) { return backend_->fetchAllTransactionsInLedger(lgrInfo.seq, yield); });
 
                 auto ledgerRange = backend_->fetchLedgerRange();
@@ -174,10 +177,10 @@ public:
                 subscriptions_->pubBookChanges(lgrInfo, transactions);
 
                 setLastPublishTime();
-                log_.info() << "Published ledger " << std::to_string(lgrInfo.seq);
+                LOG(log_.info()) << "Published ledger " << std::to_string(lgrInfo.seq);
             }
             else
-                log_.info() << "Skipping publishing ledger " << std::to_string(lgrInfo.seq);
+                LOG(log_.info()) << "Skipping publishing ledger " << std::to_string(lgrInfo.seq);
         });
 
         // we track latest publish-requested seq, not necessarily already published
@@ -249,4 +252,4 @@ private:
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/etl/impl/Transformer.h

@@ -19,22 +19,23 @@
 
 #pragma once
 
-#include <backend/BackendInterface.h>
+#include <data/BackendInterface.h>
 #include <etl/SystemState.h>
+#include <etl/impl/AmendmentBlock.h>
 #include <etl/impl/LedgerLoader.h>
-#include <log/Logger.h>
 #include <util/LedgerUtils.h>
 #include <util/Profiler.h>
+#include <util/log/Logger.h>
 
 #include <ripple/beast/core/CurrentThreadName.h>
-#include "org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h"
+#include <ripple/proto/org/xrpl/rpc/v1/xrp_ledger.grpc.pb.h>
 #include <grpcpp/grpcpp.h>
 
 #include <chrono>
 #include <memory>
 #include <thread>
 
-namespace clio::detail {
+namespace etl::detail {
 
 /*
  * TODO:
@@ -45,20 +46,26 @@ namespace clio::detail {
  */
 
 /**
- * @brief Transformer thread that prepares new ledger out of raw data from GRPC
+ * @brief Transformer thread that prepares new ledger out of raw data from GRPC.
  */
-template <typename DataPipeType, typename LedgerLoaderType, typename LedgerPublisherType>
+template <
+    typename DataPipeType,
+    typename LedgerLoaderType,
+    typename LedgerPublisherType,
+    typename AmendmentBlockHandlerType>
 class Transformer
 {
     using GetLedgerResponseType = typename LedgerLoaderType::GetLedgerResponseType;
     using RawLedgerObjectType = typename LedgerLoaderType::RawLedgerObjectType;
 
-    clio::Logger log_{"ETL"};
+    util::Logger log_{"ETL"};
 
     std::reference_wrapper<DataPipeType> pipe_;
     std::shared_ptr<BackendInterface> backend_;
     std::reference_wrapper<LedgerLoaderType> loader_;
     std::reference_wrapper<LedgerPublisherType> publisher_;
+    std::reference_wrapper<AmendmentBlockHandlerType> amendmentBlockHandler_;
+
     uint32_t startSequence_;
     std::reference_wrapper<SystemState> state_;  // shared state for ETL
 
@@ -66,7 +73,7 @@ class Transformer
 
 public:
     /**
-     * @brief Create an instance of the transformer
+     * @brief Create an instance of the transformer.
      *
      * This spawns a new thread that reads from the data pipe and writes ledgers to the DB using LedgerLoader and
      * LedgerPublisher.
@@ -76,12 +83,14 @@ public:
         std::shared_ptr<BackendInterface> backend,
         LedgerLoaderType& loader,
         LedgerPublisherType& publisher,
+        AmendmentBlockHandlerType& amendmentBlockHandler,
         uint32_t startSequence,
         SystemState& state)
-        : pipe_(std::ref(pipe))
+        : pipe_{std::ref(pipe)}
         , backend_{backend}
-        , loader_(std::ref(loader))
-        , publisher_(std::ref(publisher))
+        , loader_{std::ref(loader)}
+        , publisher_{std::ref(publisher)}
+        , amendmentBlockHandler_{std::ref(amendmentBlockHandler)}
         , startSequence_{startSequence}
         , state_{std::ref(state)}
     {
@@ -89,7 +98,7 @@ public:
     }
 
     /**
-     * @brief Joins the transformer thread
+     * @brief Joins the transformer thread.
      */
     ~Transformer()
     {
@@ -98,7 +107,7 @@ public:
     }
 
     /**
-     * @brief Block calling thread until transformer thread exits
+     * @brief Block calling thread until transformer thread exits.
      */
     void
     waitTillFinished()
@@ -137,62 +146,72 @@ private:
                 auto const end = std::chrono::system_clock::now();
                 auto const duration = ((end - start).count()) / 1000000000.0;
 
-                log_.info() << "Load phase of etl : "
-                            << "Successfully wrote ledger! Ledger info: " << util::toString(lgrInfo)
-                            << ". txn count = " << numTxns << ". object count = " << numObjects
-                            << ". load time = " << duration << ". load txns per second = " << numTxns / duration
-                            << ". load objs per second = " << numObjects / duration;
+                LOG(log_.info()) << "Load phase of etl : "
+                                 << "Successfully wrote ledger! Ledger info: " << util::toString(lgrInfo)
+                                 << ". txn count = " << numTxns << ". object count = " << numObjects
+                                 << ". load time = " << duration << ". load txns per second = " << numTxns / duration
+                                 << ". load objs per second = " << numObjects / duration;
 
                 // success is false if the ledger was already written
                 publisher_.get().publish(lgrInfo);
             }
             else
             {
-                log_.error() << "Error writing ledger. " << util::toString(lgrInfo);
+                LOG(log_.error()) << "Error writing ledger. " << util::toString(lgrInfo);
             }
 
             setWriteConflict(not success);
         }
     }
 
-    // TODO update this documentation
     /**
      * @brief Build the next ledger using the previous ledger and the extracted data.
      * @note rawData should be data that corresponds to the ledger immediately following the previous seq.
      *
-     * @param rawData data extracted from an ETL source
-     * @return the newly built ledger and data to write to the database
+     * @param rawData Data extracted from an ETL source
+     * @return The newly built ledger and data to write to the database
      */
-    std::pair<ripple::LedgerInfo, bool>
+    std::pair<ripple::LedgerHeader, bool>
     buildNextLedger(GetLedgerResponseType& rawData)
     {
-        log_.debug() << "Beginning ledger update";
-        ripple::LedgerInfo lgrInfo = util::deserializeHeader(ripple::makeSlice(rawData.ledger_header()));
+        LOG(log_.debug()) << "Beginning ledger update";
+        ripple::LedgerHeader lgrInfo = ::util::deserializeHeader(ripple::makeSlice(rawData.ledger_header()));
 
-        log_.debug() << "Deserialized ledger header. " << util::toString(lgrInfo);
+        LOG(log_.debug()) << "Deserialized ledger header. " << ::util::toString(lgrInfo);
         backend_->startWrites();
         backend_->writeLedger(lgrInfo, std::move(*rawData.mutable_ledger_header()));
 
         writeSuccessors(lgrInfo, rawData);
-        updateCache(lgrInfo, rawData);
+        std::optional<FormattedTransactionsData> insertTxResultOp;
+        try
+        {
+            updateCache(lgrInfo, rawData);
 
-        log_.debug() << "Inserted/modified/deleted all objects. Number of objects = "
-                     << rawData.ledger_objects().objects_size();
+            LOG(log_.debug()) << "Inserted/modified/deleted all objects. Number of objects = "
+                              << rawData.ledger_objects().objects_size();
 
-        auto insertTxResult = loader_.get().insertTransactions(lgrInfo, rawData);
+            insertTxResultOp.emplace(loader_.get().insertTransactions(lgrInfo, rawData));
+        }
+        catch (std::runtime_error const& e)
+        {
+            LOG(log_.fatal()) << "Failed to build next ledger: " << e.what();
 
-        log_.debug() << "Inserted all transactions. Number of transactions  = "
-                     << rawData.transactions_list().transactions_size();
+            amendmentBlockHandler_.get().onAmendmentBlock();
+            return {ripple::LedgerHeader{}, false};
+        }
 
-        backend_->writeAccountTransactions(std::move(insertTxResult.accountTxData));
-        backend_->writeNFTs(std::move(insertTxResult.nfTokensData));
-        backend_->writeNFTTransactions(std::move(insertTxResult.nfTokenTxData));
+        LOG(log_.debug()) << "Inserted all transactions. Number of transactions  = "
+                          << rawData.transactions_list().transactions_size();
+
+        backend_->writeAccountTransactions(std::move(insertTxResultOp->accountTxData));
+        backend_->writeNFTs(std::move(insertTxResultOp->nfTokensData));
+        backend_->writeNFTTransactions(std::move(insertTxResultOp->nfTokenTxData));
 
         auto [success, duration] =
-            util::timed<std::chrono::duration<double>>([&]() { return backend_->finishWrites(lgrInfo.seq); });
+            ::util::timed<std::chrono::duration<double>>([&]() { return backend_->finishWrites(lgrInfo.seq); });
 
-        log_.debug() << "Finished writes. Total time: " << std::to_string(duration);
-        log_.debug() << "Finished ledger update: " << util::toString(lgrInfo);
+        LOG(log_.debug()) << "Finished writes. Total time: " << std::to_string(duration);
+        LOG(log_.debug()) << "Finished ledger update: " << ::util::toString(lgrInfo);
 
         return {lgrInfo, success};
     }
@@ -204,9 +223,9 @@ private:
      * @param rawData Ledger data from GRPC
      */
     void
-    updateCache(ripple::LedgerInfo const& lgrInfo, GetLedgerResponseType& rawData)
+    updateCache(ripple::LedgerHeader const& lgrInfo, GetLedgerResponseType& rawData)
     {
-        std::vector<Backend::LedgerObject> cacheUpdates;
+        std::vector<data::LedgerObject> cacheUpdates;
         cacheUpdates.reserve(rawData.ledger_objects().objects_size());
 
         // TODO change these to unordered_set
@@ -219,14 +238,14 @@ private:
             assert(key);
 
             cacheUpdates.push_back({*key, {obj.mutable_data()->begin(), obj.mutable_data()->end()}});
-            log_.debug() << "key = " << ripple::strHex(*key) << " - mod type = " << obj.mod_type();
+            LOG(log_.debug()) << "key = " << ripple::strHex(*key) << " - mod type = " << obj.mod_type();
 
             if (obj.mod_type() != RawLedgerObjectType::MODIFIED && !rawData.object_neighbors_included())
             {
-                log_.debug() << "object neighbors not included. using cache";
+                LOG(log_.debug()) << "object neighbors not included. using cache";
 
                 if (!backend_->cache().isFull() || backend_->cache().latestLedgerSequence() != lgrInfo.seq - 1)
-                    throw std::runtime_error("Cache is not full, but object neighbors were not included");
+                    throw std::logic_error("Cache is not full, but object neighbors were not included");
 
                 auto const blob = obj.mutable_data();
                 auto checkBookBase = false;
@@ -245,7 +264,7 @@ private:
 
                 if (checkBookBase)
                 {
-                    log_.debug() << "Is book dir. Key = " << ripple::strHex(*key);
+                    LOG(log_.debug()) << "Is book dir. Key = " << ripple::strHex(*key);
 
                     auto const bookBase = getBookBase(*key);
                     auto const oldFirstDir = backend_->cache().getSuccessor(bookBase, lgrInfo.seq - 1);
@@ -254,9 +273,10 @@ private:
                     // We deleted the first directory, or we added a directory prior to the old first directory
                     if ((isDeleted && key == oldFirstDir->key) || (!isDeleted && key < oldFirstDir->key))
                     {
-                        log_.debug() << "Need to recalculate book base successor. base = " << ripple::strHex(bookBase)
-                                     << " - key = " << ripple::strHex(*key) << " - isDeleted = " << isDeleted
-                                     << " - seq = " << lgrInfo.seq;
+                        LOG(log_.debug())
+                            << "Need to recalculate book base successor. base = " << ripple::strHex(bookBase)
+                            << " - key = " << ripple::strHex(*key) << " - isDeleted = " << isDeleted
+                            << " - seq = " << lgrInfo.seq;
                         bookSuccessorsToCalculate.insert(bookBase);
                     }
                 }
@@ -273,9 +293,9 @@ private:
         // rippled didn't send successor information, so use our cache
         if (!rawData.object_neighbors_included())
         {
-            log_.debug() << "object neighbors not included. using cache";
+            LOG(log_.debug()) << "object neighbors not included. using cache";
             if (!backend_->cache().isFull() || backend_->cache().latestLedgerSequence() != lgrInfo.seq)
-                throw std::runtime_error("Cache is not full, but object neighbors were not included");
+                throw std::logic_error("Cache is not full, but object neighbors were not included");
 
             for (auto const& obj : cacheUpdates)
             {
@@ -284,16 +304,16 @@ private:
 
                 auto lb = backend_->cache().getPredecessor(obj.key, lgrInfo.seq);
                 if (!lb)
-                    lb = {Backend::firstKey, {}};
+                    lb = {data::firstKey, {}};
 
                 auto ub = backend_->cache().getSuccessor(obj.key, lgrInfo.seq);
                 if (!ub)
-                    ub = {Backend::lastKey, {}};
+                    ub = {data::lastKey, {}};
 
                 if (obj.blob.size() == 0)
                 {
-                    log_.debug() << "writing successor for deleted object " << ripple::strHex(obj.key) << " - "
-                                 << ripple::strHex(lb->key) << " - " << ripple::strHex(ub->key);
+                    LOG(log_.debug()) << "writing successor for deleted object " << ripple::strHex(obj.key) << " - "
+                                      << ripple::strHex(lb->key) << " - " << ripple::strHex(ub->key);
 
                     backend_->writeSuccessor(uint256ToString(lb->key), lgrInfo.seq, uint256ToString(ub->key));
                 }
@@ -302,8 +322,8 @@ private:
                     backend_->writeSuccessor(uint256ToString(lb->key), lgrInfo.seq, uint256ToString(obj.key));
                     backend_->writeSuccessor(uint256ToString(obj.key), lgrInfo.seq, uint256ToString(ub->key));
 
-                    log_.debug() << "writing successor for new object " << ripple::strHex(lb->key) << " - "
-                                 << ripple::strHex(obj.key) << " - " << ripple::strHex(ub->key);
+                    LOG(log_.debug()) << "writing successor for new object " << ripple::strHex(lb->key) << " - "
+                                      << ripple::strHex(obj.key) << " - " << ripple::strHex(ub->key);
                 }
             }
 
@@ -314,41 +334,41 @@ private:
                 {
                     backend_->writeSuccessor(uint256ToString(base), lgrInfo.seq, uint256ToString(succ->key));
 
-                    log_.debug() << "Updating book successor " << ripple::strHex(base) << " - "
-                                 << ripple::strHex(succ->key);
+                    LOG(log_.debug()) << "Updating book successor " << ripple::strHex(base) << " - "
+                                      << ripple::strHex(succ->key);
                 }
                 else
                 {
-                    backend_->writeSuccessor(uint256ToString(base), lgrInfo.seq, uint256ToString(Backend::lastKey));
+                    backend_->writeSuccessor(uint256ToString(base), lgrInfo.seq, uint256ToString(data::lastKey));
 
-                    log_.debug() << "Updating book successor " << ripple::strHex(base) << " - "
-                                 << ripple::strHex(Backend::lastKey);
+                    LOG(log_.debug()) << "Updating book successor " << ripple::strHex(base) << " - "
+                                      << ripple::strHex(data::lastKey);
                 }
             }
         }
     }
 
     /**
-     * @brief Write successors info into DB
+     * @brief Write successors info into DB.
      *
      * @param lgrInfo Ledger info
      * @param rawData Ledger data from GRPC
      */
     void
-    writeSuccessors(ripple::LedgerInfo const& lgrInfo, GetLedgerResponseType& rawData)
+    writeSuccessors(ripple::LedgerHeader const& lgrInfo, GetLedgerResponseType& rawData)
     {
         // Write successor info, if included from rippled
         if (rawData.object_neighbors_included())
         {
-            log_.debug() << "object neighbors included";
+            LOG(log_.debug()) << "object neighbors included";
 
             for (auto& obj : *(rawData.mutable_book_successors()))
             {
                 auto firstBook = std::move(*obj.mutable_first_book());
                 if (!firstBook.size())
-                    firstBook = uint256ToString(Backend::lastKey);
-                log_.debug() << "writing book successor " << ripple::strHex(obj.book_base()) << " - "
-                             << ripple::strHex(firstBook);
+                    firstBook = uint256ToString(data::lastKey);
+                LOG(log_.debug()) << "writing book successor " << ripple::strHex(obj.book_base()) << " - "
+                                  << ripple::strHex(firstBook);
 
                 backend_->writeSuccessor(std::move(*obj.mutable_book_base()), lgrInfo.seq, std::move(firstBook));
             }
@@ -359,45 +379,52 @@ private:
                 {
                     std::string* predPtr = obj.mutable_predecessor();
                     if (!predPtr->size())
-                        *predPtr = uint256ToString(Backend::firstKey);
+                        *predPtr = uint256ToString(data::firstKey);
                     std::string* succPtr = obj.mutable_successor();
                     if (!succPtr->size())
-                        *succPtr = uint256ToString(Backend::lastKey);
+                        *succPtr = uint256ToString(data::lastKey);
 
                     if (obj.mod_type() == RawLedgerObjectType::DELETED)
                     {
-                        log_.debug() << "Modifying successors for deleted object " << ripple::strHex(obj.key()) << " - "
-                                     << ripple::strHex(*predPtr) << " - " << ripple::strHex(*succPtr);
+                        LOG(log_.debug()) << "Modifying successors for deleted object " << ripple::strHex(obj.key())
+                                          << " - " << ripple::strHex(*predPtr) << " - " << ripple::strHex(*succPtr);
 
                         backend_->writeSuccessor(std::move(*predPtr), lgrInfo.seq, std::move(*succPtr));
                     }
                     else
                     {
-                        log_.debug() << "adding successor for new object " << ripple::strHex(obj.key()) << " - "
-                                     << ripple::strHex(*predPtr) << " - " << ripple::strHex(*succPtr);
+                        LOG(log_.debug()) << "adding successor for new object " << ripple::strHex(obj.key()) << " - "
+                                          << ripple::strHex(*predPtr) << " - " << ripple::strHex(*succPtr);
 
                         backend_->writeSuccessor(std::move(*predPtr), lgrInfo.seq, std::string{obj.key()});
                         backend_->writeSuccessor(std::string{obj.key()}, lgrInfo.seq, std::move(*succPtr));
                     }
                 }
                 else
-                    log_.debug() << "object modified " << ripple::strHex(obj.key());
+                    LOG(log_.debug()) << "object modified " << ripple::strHex(obj.key());
             }
         }
     }
 
+    /** @return true if the transformer is stopping; false otherwise */
     bool
     isStopping() const
     {
         return state_.get().isStopping;
     }
 
+    /** @return true if there was a write conflict; false otherwise */
     bool
     hasWriteConflict() const
     {
         return state_.get().writeConflict;
     }
 
+    /**
+     * @brief Sets the write conflict flag.
+     *
+     * @param conflict The value to set
+     */
     void
     setWriteConflict(bool conflict)
     {
@@ -405,4 +432,4 @@ private:
     }
 };
 
-}  // namespace clio::detail
+}  // namespace etl::detail

src/feed/SubscriptionManager.cpp

@@ -17,9 +17,11 @@
 */
 //==============================================================================
 
+#include <feed/SubscriptionManager.h>
 #include <rpc/BookChangesHelper.h>
 #include <rpc/RPCHelpers.h>
-#include <subscriptions/SubscriptionManager.h>
+
+namespace feed {
 
 void
 Subscription::subscribe(SessionPtrType const& session)
@@ -41,7 +43,7 @@ Subscription::publish(std::shared_ptr<std::string> const& message)
 
 boost::json::object
 getLedgerPubMessage(
-    ripple::LedgerInfo const& lgrInfo,
+    ripple::LedgerHeader const& lgrInfo,
     ripple::Fees const& fees,
     std::string const& ledgerRange,
     std::uint32_t txnCount)
@@ -53,10 +55,9 @@ getLedgerPubMessage(
     pubMsg["ledger_hash"] = to_string(lgrInfo.hash);
     pubMsg["ledger_time"] = lgrInfo.closeTime.time_since_epoch().count();
 
-    pubMsg["fee_ref"] = RPC::toBoostJson(fees.units.jsonClipped());
-    pubMsg["fee_base"] = RPC::toBoostJson(fees.base.jsonClipped());
-    pubMsg["reserve_base"] = RPC::toBoostJson(fees.reserve.jsonClipped());
-    pubMsg["reserve_inc"] = RPC::toBoostJson(fees.increment.jsonClipped());
+    pubMsg["fee_base"] = rpc::toBoostJson(fees.base.jsonClipped());
+    pubMsg["reserve_base"] = rpc::toBoostJson(fees.reserve.jsonClipped());
+    pubMsg["reserve_inc"] = rpc::toBoostJson(fees.increment.jsonClipped());
 
     pubMsg["validated_ledgers"] = ledgerRange;
     pubMsg["txn_count"] = txnCount;
@@ -64,7 +65,7 @@ getLedgerPubMessage(
 }
 
 boost::json::object
-SubscriptionManager::subLedger(boost::asio::yield_context& yield, SessionPtrType session)
+SubscriptionManager::subLedger(boost::asio::yield_context yield, SessionPtrType session)
 {
     subscribeHelper(session, ledgerSubscribers_, [this](SessionPtrType session) { unsubLedger(session); });
 
@@ -144,7 +145,7 @@ SubscriptionManager::unsubBookChanges(SessionPtrType session)
 
 void
 SubscriptionManager::pubLedger(
-    ripple::LedgerInfo const& lgrInfo,
+    ripple::LedgerHeader const& lgrInfo,
     ripple::Fees const& fees,
     std::string const& ledgerRange,
     std::uint32_t txnCount)
@@ -156,13 +157,13 @@ SubscriptionManager::pubLedger(
 }
 
 void
-SubscriptionManager::pubTransaction(Backend::TransactionAndMetadata const& blobs, ripple::LedgerInfo const& lgrInfo)
+SubscriptionManager::pubTransaction(data::TransactionAndMetadata const& blobs, ripple::LedgerHeader const& lgrInfo)
 {
-    auto [tx, meta] = RPC::deserializeTxPlusMeta(blobs, lgrInfo.seq);
+    auto [tx, meta] = rpc::deserializeTxPlusMeta(blobs, lgrInfo.seq);
     boost::json::object pubObj;
-    pubObj["transaction"] = RPC::toJson(*tx);
-    pubObj["meta"] = RPC::toJson(*meta);
-    RPC::insertDeliveredAmount(pubObj["meta"].as_object(), tx, meta, blobs.date);
+    pubObj["transaction"] = rpc::toJson(*tx);
+    pubObj["meta"] = rpc::toJson(*meta);
+    rpc::insertDeliveredAmount(pubObj["meta"].as_object(), tx, meta, blobs.date);
     pubObj["type"] = "transaction";
     pubObj["validated"] = true;
     pubObj["status"] = "closed";
@@ -185,12 +186,12 @@ SubscriptionManager::pubTransaction(Backend::TransactionAndMetadata const& blobs
         {
             ripple::STAmount ownerFunds;
             auto fetchFundsSynchronous = [&]() {
-                Backend::synchronous([&](boost::asio::yield_context& yield) {
-                    ownerFunds = RPC::accountFunds(*backend_, lgrInfo.seq, amount, account, yield);
+                data::synchronous([&](boost::asio::yield_context yield) {
+                    ownerFunds = rpc::accountFunds(*backend_, lgrInfo.seq, amount, account, yield);
                 });
             };
 
-            Backend::retryOnTimeout(fetchFundsSynchronous);
+            data::retryOnTimeout(fetchFundsSynchronous);
 
             pubObj["transaction"].as_object()["owner_funds"] = ownerFunds.getText();
         }
@@ -244,10 +245,10 @@ SubscriptionManager::pubTransaction(Backend::TransactionAndMetadata const& blobs
 
 void
 SubscriptionManager::pubBookChanges(
-    ripple::LedgerInfo const& lgrInfo,
-    std::vector<Backend::TransactionAndMetadata> const& transactions)
+    ripple::LedgerHeader const& lgrInfo,
+    std::vector<data::TransactionAndMetadata> const& transactions)
 {
-    auto const json = RPC::computeBookChanges(lgrInfo, transactions);
+    auto const json = rpc::computeBookChanges(lgrInfo, transactions);
     auto const bookChangesMsg = std::make_shared<std::string>(boost::json::serialize(json));
     bookChangesSubscribers_.publish(bookChangesMsg);
 }
@@ -259,7 +260,7 @@ SubscriptionManager::forwardProposedTransaction(boost::json::object const& respo
     txProposedSubscribers_.publish(pubMsg);
 
     auto transaction = response.at("transaction").as_object();
-    auto accounts = RPC::getAccountsFromTransaction(transaction);
+    auto accounts = rpc::getAccountsFromTransaction(transaction);
 
     for (ripple::AccountID const& account : accounts)
         accountProposedSubscribers_.publish(pubMsg, account);
@@ -365,3 +366,5 @@ SubscriptionManager::cleanup(SessionPtrType session)
 
     cleanupFuncs_.erase(session);
 }
+
+}  // namespace feed

src/feed/SubscriptionManager.h

@@ -0,0 +1,591 @@
+//------------------------------------------------------------------------------
+/*
+    This file is part of clio: https://github.com/XRPLF/clio
+    Copyright (c) 2022, the clio developers.
+
+    Permission to use, copy, modify, and distribute this software for any
+    purpose with or without fee is hereby granted, provided that the above
+    copyright notice and this permission notice appear in all copies.
+
+    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
+    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
+    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+//==============================================================================
+
+#pragma once
+
+#include <data/BackendInterface.h>
+#include <util/config/Config.h>
+#include <util/log/Logger.h>
+#include <web/interface/ConnectionBase.h>
+
+#include <ripple/protocol/LedgerHeader.h>
+
+#include <memory>
+
+/**
+ * @brief This namespace deals with subscriptions.
+ */
+namespace feed {
+
+using SessionPtrType = std::shared_ptr<web::ConnectionBase>;
+
+/**
+ * @brief Sends a message to subscribers.
+ *
+ * @param message The message to send
+ * @param subscribers The subscription stream to send the message to
+ * @param counter The subscription counter to decrement if session is detected as dead
+ */
+template <class T>
+inline void
+sendToSubscribers(std::shared_ptr<std::string> const& message, T& subscribers, std::atomic_uint64_t& counter)
+{
+    for (auto it = subscribers.begin(); it != subscribers.end();)
+    {
+        auto& session = *it;
+        if (session->dead())
+        {
+            it = subscribers.erase(it);
+            --counter;
+        }
+        else
+        {
+            session->send(message);
+            ++it;
+        }
+    }
+}
+
+/**
+ * @brief Adds a session to the subscription stream.
+ *
+ * @param session The session to add
+ * @param subscribers The stream to subscribe to
+ * @param counter The counter representing the current total subscribers
+ */
+template <class T>
+inline void
+addSession(SessionPtrType session, T& subscribers, std::atomic_uint64_t& counter)
+{
+    if (!subscribers.contains(session))
+    {
+        subscribers.insert(session);
+        ++counter;
+    }
+}
+
+/**
+ * @brief Removes a session from the subscription stream.
+ *
+ * @param session The session to remove
+ * @param subscribers The stream to unsubscribe from
+ * @param counter The counter representing the current total subscribers
+ */
+template <class T>
+inline void
+removeSession(SessionPtrType session, T& subscribers, std::atomic_uint64_t& counter)
+{
+    if (subscribers.contains(session))
+    {
+        subscribers.erase(session);
+        --counter;
+    }
+}
+
+/**
+ * @brief Represents a subscription stream.
+ */
+class Subscription
+{
+    boost::asio::strand<boost::asio::io_context::executor_type> strand_;
+    std::unordered_set<SessionPtrType> subscribers_ = {};
+    std::atomic_uint64_t subCount_ = 0;
+
+public:
+    Subscription() = delete;
+    Subscription(Subscription&) = delete;
+    Subscription(Subscription&&) = delete;
+
+    /**
+     * @brief Create a new subscription stream.
+     *
+     * @param ioc The io_context to run on
+     */
+    explicit Subscription(boost::asio::io_context& ioc) : strand_(boost::asio::make_strand(ioc))
+    {
+    }
+
+    ~Subscription() = default;
+
+    /**
+     * @brief Adds the given session to the subscribers set.
+     *
+     * @param session The session to add
+     */
+    void
+    subscribe(SessionPtrType const& session);
+
+    /**
+     * @brief Removes the given session from the subscribers set.
+     *
+     * @param session The session to remove
+     */
+    void
+    unsubscribe(SessionPtrType const& session);
+
+    /**
+     * @brief Sends the given message to all subscribers.
+     *
+     * @param message The message to send
+     */
+    void
+    publish(std::shared_ptr<std::string> const& message);
+
+    /**
+     * @return Total subscriber count on this stream.
+     */
+    std::uint64_t
+    count() const
+    {
+        return subCount_.load();
+    }
+
+    /**
+     * @return true if the stream currently has no subscribers; false otherwise
+     */
+    bool
+    empty() const
+    {
+        return count() == 0;
+    }
+};
+
+/**
+ * @brief Represents a collection of subscriptions where each stream is mapped to a key.
+ */
+template <class Key>
+class SubscriptionMap
+{
+    using SubscribersType = std::set<SessionPtrType>;
+
+    boost::asio::strand<boost::asio::io_context::executor_type> strand_;
+    std::unordered_map<Key, SubscribersType> subscribers_ = {};
+    std::atomic_uint64_t subCount_ = 0;
+
+public:
+    SubscriptionMap() = delete;
+    SubscriptionMap(SubscriptionMap&) = delete;
+    SubscriptionMap(SubscriptionMap&&) = delete;
+
+    /**
+     * @brief Create a new subscription map.
+     *
+     * @param ioc The io_context to run on
+     */
+    explicit SubscriptionMap(boost::asio::io_context& ioc) : strand_(boost::asio::make_strand(ioc))
+    {
+    }
+
+    ~SubscriptionMap() = default;
+
+    /**
+     * @brief Subscribe to a specific stream by its key.
+     *
+     * @param session The session to add
+     * @param key The key for the subscription to subscribe to
+     */
+    void
+    subscribe(SessionPtrType const& session, Key const& key)
+    {
+        boost::asio::post(strand_, [this, session, key]() { addSession(session, subscribers_[key], subCount_); });
+    }
+
+    /**
+     * @brief Unsubscribe from a specific stream by its key.
+     *
+     * @param session The session to remove
+     * @param key The key for the subscription to unsubscribe from
+     */
+    void
+    unsubscribe(SessionPtrType const& session, Key const& key)
+    {
+        boost::asio::post(strand_, [this, key, session]() {
+            if (!subscribers_.contains(key))
+                return;
+
+            if (!subscribers_[key].contains(session))
+                return;
+
+            --subCount_;
+            subscribers_[key].erase(session);
+
+            if (subscribers_[key].size() == 0)
+            {
+                subscribers_.erase(key);
+            }
+        });
+    }
+
+    /**
+     * @brief Sends the given message to all subscribers.
+     *
+     * @param message The message to send
+     * @param key The key for the subscription to send the message to
+     */
+    void
+    publish(std::shared_ptr<std::string> const& message, Key const& key)
+    {
+        boost::asio::post(strand_, [this, key, message]() {
+            if (!subscribers_.contains(key))
+                return;
+
+            sendToSubscribers(message, subscribers_[key], subCount_);
+        });
+    }
+
+    /**
+     * @return Total subscriber count on all streams in the collection.
+     */
+    std::uint64_t
+    count() const
+    {
+        return subCount_.load();
+    }
+};
+
+/**
+ * @brief Manages subscriptions.
+ */
+class SubscriptionManager
+{
+    util::Logger log_{"Subscriptions"};
+
+    std::vector<std::thread> workers_;
+    boost::asio::io_context ioc_;
+    std::optional<boost::asio::io_context::work> work_;
+
+    Subscription ledgerSubscribers_;
+    Subscription txSubscribers_;
+    Subscription txProposedSubscribers_;
+    Subscription manifestSubscribers_;
+    Subscription validationsSubscribers_;
+    Subscription bookChangesSubscribers_;
+
+    SubscriptionMap<ripple::AccountID> accountSubscribers_;
+    SubscriptionMap<ripple::AccountID> accountProposedSubscribers_;
+    SubscriptionMap<ripple::Book> bookSubscribers_;
+
+    std::shared_ptr<data::BackendInterface const> backend_;
+
+public:
+    /**
+     * @brief A factory function that creates a new subscription manager configured from the config provided.
+     *
+     * @param config The configuration to use
+     * @param backend The backend to use
+     */
+    static std::shared_ptr<SubscriptionManager>
+    make_SubscriptionManager(util::Config const& config, std::shared_ptr<data::BackendInterface const> const& backend)
+    {
+        auto numThreads = config.valueOr<uint64_t>("subscription_workers", 1);
+        return std::make_shared<SubscriptionManager>(numThreads, backend);
+    }
+
+    /**
+     * @brief Creates a new instance of the subscription manager.
+     *
+     * @param numThreads The number of worker threads to manage subscriptions
+     * @param backend The backend to use
+     */
+    SubscriptionManager(std::uint64_t numThreads, std::shared_ptr<data::BackendInterface const> const& backend)
+        : ledgerSubscribers_(ioc_)
+        , txSubscribers_(ioc_)
+        , txProposedSubscribers_(ioc_)
+        , manifestSubscribers_(ioc_)
+        , validationsSubscribers_(ioc_)
+        , bookChangesSubscribers_(ioc_)
+        , accountSubscribers_(ioc_)
+        , accountProposedSubscribers_(ioc_)
+        , bookSubscribers_(ioc_)
+        , backend_(backend)
+    {
+        work_.emplace(ioc_);
+
+        // We will eventually want to clamp this to be the number of strands,
+        // since adding more threads than we have strands won't see any
+        // performance benefits
+        LOG(log_.info()) << "Starting subscription manager with " << numThreads << " workers";
+
+        workers_.reserve(numThreads);
+        for (auto i = numThreads; i > 0; --i)
+            workers_.emplace_back([this] { ioc_.run(); });
+    }
+
+    /** @brief Stops the worker threads of the subscription manager. */
+    ~SubscriptionManager()
+    {
+        work_.reset();
+
+        ioc_.stop();
+        for (auto& worker : workers_)
+            worker.join();
+    }
+
+    /**
+     * @brief Subscribe to the ledger stream.
+     *
+     * @param yield The coroutine context
+     * @param session The session to subscribe to the stream
+     * @return JSON object representing the first message to be sent to the new subscriber
+     */
+    boost::json::object
+    subLedger(boost::asio::yield_context yield, SessionPtrType session);
+
+    /**
+     * @brief Publish to the ledger stream.
+     *
+     * @param lgrInfo The ledger header to serialize
+     * @param fees The fees to serialize
+     * @param ledgerRange The ledger range this message applies to
+     * @param txnCount The total number of transactions to serialize
+     */
+    void
+    pubLedger(
+        ripple::LedgerHeader const& lgrInfo,
+        ripple::Fees const& fees,
+        std::string const& ledgerRange,
+        std::uint32_t txnCount);
+
+    /**
+     * @brief Publish to the book changes stream.
+     *
+     * @param lgrInfo The ledger header to serialize
+     * @param transactions The transactions to serialize
+     */
+    void
+    pubBookChanges(ripple::LedgerHeader const& lgrInfo, std::vector<data::TransactionAndMetadata> const& transactions);
+
+    /**
+     * @brief Unsubscribe from the ledger stream.
+     *
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubLedger(SessionPtrType session);
+
+    /**
+     * @brief Subscribe to the transactions stream.
+     *
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subTransactions(SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the transactions stream.
+     *
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubTransactions(SessionPtrType session);
+
+    /**
+     * @brief Publish to the book changes stream.
+     *
+     * @param blobs The transactions to serialize
+     * @param lgrInfo The ledger header to serialize
+     */
+    void
+    pubTransaction(data::TransactionAndMetadata const& blobs, ripple::LedgerHeader const& lgrInfo);
+
+    /**
+     * @brief Subscribe to the account changes stream.
+     *
+     * @param account The account to monitor changes for
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subAccount(ripple::AccountID const& account, SessionPtrType const& session);
+
+    /**
+     * @brief Unsubscribe from the account changes stream.
+     *
+     * @param account The account the stream is for
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubAccount(ripple::AccountID const& account, SessionPtrType const& session);
+
+    /**
+     * @brief Subscribe to a specific book changes stream.
+     *
+     * @param book The book to monitor changes for
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subBook(ripple::Book const& book, SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the specific book changes stream.
+     *
+     * @param book The book to stop monitoring changes for
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubBook(ripple::Book const& book, SessionPtrType session);
+
+    /**
+     * @brief Subscribe to the book changes stream.
+     *
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subBookChanges(SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the book changes stream.
+     *
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubBookChanges(SessionPtrType session);
+
+    /**
+     * @brief Subscribe to the manifest stream.
+     *
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subManifest(SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the manifest stream.
+     *
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubManifest(SessionPtrType session);
+
+    /**
+     * @brief Subscribe to the validation stream.
+     *
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subValidation(SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the validation stream.
+     *
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubValidation(SessionPtrType session);
+
+    /**
+     * @brief Publish proposed transactions and proposed accounts from a JSON response.
+     *
+     * @param response The JSON response to use
+     */
+    void
+    forwardProposedTransaction(boost::json::object const& response);
+
+    /**
+     * @brief Publish manifest updates from a JSON response.
+     *
+     * @param response The JSON response to use
+     */
+    void
+    forwardManifest(boost::json::object const& response);
+
+    /**
+     * @brief Publish validation updates from a JSON response.
+     *
+     * @param response The JSON response to use
+     */
+    void
+    forwardValidation(boost::json::object const& response);
+
+    /**
+     * @brief Subscribe to the proposed account stream.
+     *
+     * @param account The account to monitor
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subProposedAccount(ripple::AccountID const& account, SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the proposed account stream.
+     *
+     * @param account The account the stream is for
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubProposedAccount(ripple::AccountID const& account, SessionPtrType session);
+
+    /**
+     * @brief Subscribe to the processed transactions stream.
+     *
+     * @param session The session to subscribe to the stream
+     */
+    void
+    subProposedTransactions(SessionPtrType session);
+
+    /**
+     * @brief Unsubscribe from the proposed transactions stream.
+     *
+     * @param session The session to unsubscribe from the stream
+     */
+    void
+    unsubProposedTransactions(SessionPtrType session);
+
+    /** @brief Clenup the session on removal. */
+    void
+    cleanup(SessionPtrType session);
+
+    /**
+     * @brief Generate a JSON report on the current state of the subscriptions.
+     *
+     * @return The report as a JSON object
+     */
+    boost::json::object
+    report() const
+    {
+        return {
+            {"ledger", ledgerSubscribers_.count()},
+            {"transactions", txSubscribers_.count()},
+            {"transactions_proposed", txProposedSubscribers_.count()},
+            {"manifests", manifestSubscribers_.count()},
+            {"validations", validationsSubscribers_.count()},
+            {"account", accountSubscribers_.count()},
+            {"accounts_proposed", accountProposedSubscribers_.count()},
+            {"books", bookSubscribers_.count()},
+            {"book_changes", bookChangesSubscribers_.count()},
+        };
+    }
+
+private:
+    using CleanupFunction = std::function<void(SessionPtrType const)>;
+
+    void
+    subscribeHelper(SessionPtrType const& session, Subscription& subs, CleanupFunction&& func);
+
+    template <typename Key>
+    void
+    subscribeHelper(SessionPtrType const& session, Key const& k, SubscriptionMap<Key>& subs, CleanupFunction&& func);
+
+    // This is how we chose to cleanup subscriptions that have been closed.
+    // Each time we add a subscriber, we add the opposite lambda that unsubscribes that subscriber when cleanup is
+    // called with the session that closed.
+    std::mutex cleanupMtx_;
+    std::unordered_map<SessionPtrType, std::vector<CleanupFunction>> cleanupFuncs_ = {};
+};
+
+}  // namespace feed

src/main/Main.cpp

@@ -25,20 +25,21 @@
 #undef GRPC_ASAN_ENABLED
 #endif
 
-#include <backend/BackendFactory.h>
-#include <config/Config.h>
+#include <data/BackendFactory.h>
 #include <etl/ETLService.h>
 #include <rpc/Counters.h>
 #include <rpc/RPCEngine.h>
 #include <rpc/common/impl/HandlerProvider.h>
-#include <webserver/RPCExecutor.h>
-#include <webserver/Server.h>
+#include <util/config/Config.h>
+#include <web/RPCServerHandler.h>
+#include <web/Server.h>
 
 #include <boost/date_time/posix_time/posix_time_types.hpp>
 #include <boost/filesystem/path.hpp>
 #include <boost/json.hpp>
 #include <boost/program_options.hpp>
 
+#include <fstream>
 #include <main/Build.h>
 #include <memory>
 #include <sstream>
@@ -46,8 +47,9 @@
 #include <thread>
 #include <vector>
 
-using namespace clio;
+using namespace util;
 using namespace boost::asio;
+
 namespace po = boost::program_options;
 
 /**
@@ -162,55 +164,56 @@ try
     }
 
     LogService::init(config);
-    LogService::info() << "Clio version: " << Build::getClioFullVersionString();
+    LOG(LogService::info()) << "Clio version: " << Build::getClioFullVersionString();
 
     auto const threads = config.valueOr("io_threads", 2);
     if (threads <= 0)
     {
-        LogService::fatal() << "io_threads is less than 0";
+        LOG(LogService::fatal()) << "io_threads is less than 1";
         return EXIT_FAILURE;
     }
-    LogService::info() << "Number of io threads = " << threads;
+    LOG(LogService::info()) << "Number of io threads = " << threads;
 
     // IO context to handle all incoming requests, as well as other things.
     // This is not the only io context in the application.
     io_context ioc{threads};
 
     // Rate limiter, to prevent abuse
-    auto sweepHandler = IntervalSweepHandler{config, ioc};
-    auto dosGuard = DOSGuard{config, sweepHandler};
+    auto sweepHandler = web::IntervalSweepHandler{config, ioc};
+    auto whitelistHandler = web::WhitelistHandler{config};
+    auto dosGuard = web::DOSGuard{config, whitelistHandler, sweepHandler};
 
     // Interface to the database
-    auto backend = Backend::make_Backend(ioc, config);
+    auto backend = data::make_Backend(config);
 
     // Manages clients subscribed to streams
-    auto subscriptions = SubscriptionManager::make_SubscriptionManager(config, backend);
+    auto subscriptions = feed::SubscriptionManager::make_SubscriptionManager(config, backend);
 
     // Tracks which ledgers have been validated by the network
-    auto ledgers = NetworkValidatedLedgers::make_ValidatedLedgers();
+    auto ledgers = etl::NetworkValidatedLedgers::make_ValidatedLedgers();
 
     // Handles the connection to one or more rippled nodes.
     // ETL uses the balancer to extract data.
     // The server uses the balancer to forward RPCs to a rippled node.
     // The balancer itself publishes to streams (transactions_proposed and accounts_proposed)
-    auto balancer = LoadBalancer::make_LoadBalancer(config, ioc, backend, subscriptions, ledgers);
+    auto balancer = etl::LoadBalancer::make_LoadBalancer(config, ioc, backend, subscriptions, ledgers);
 
     // ETL is responsible for writing and publishing to streams. In read-only mode, ETL only publishes
-    auto etl = ETLService::make_ETLService(config, ioc, backend, subscriptions, balancer, ledgers);
+    auto etl = etl::ETLService::make_ETLService(config, ioc, backend, subscriptions, balancer, ledgers);
 
-    auto workQueue = WorkQueue::make_WorkQueue(config);
-    auto counters = RPC::Counters::make_Counters(workQueue);
-    auto const handlerProvider =
-        std::make_shared<RPC::detail::ProductionHandlerProvider const>(backend, subscriptions, balancer, etl, counters);
-    auto const rpcEngine = RPC::RPCEngine::make_RPCEngine(
-        config, backend, subscriptions, balancer, etl, dosGuard, workQueue, counters, handlerProvider);
+    auto workQueue = rpc::WorkQueue::make_WorkQueue(config);
+    auto counters = rpc::Counters::make_Counters(workQueue);
+    auto const handlerProvider = std::make_shared<rpc::detail::ProductionHandlerProvider const>(
+        config, backend, subscriptions, balancer, etl, counters);
+    auto const rpcEngine = rpc::RPCEngine::make_RPCEngine(
+        backend, subscriptions, balancer, dosGuard, workQueue, counters, handlerProvider);
 
-    // init the web server
-    auto executor =
-        std::make_shared<RPCExecutor<RPC::RPCEngine, ETLService>>(config, backend, rpcEngine, etl, subscriptions);
+    // Init the web server
+    auto handler = std::make_shared<web::RPCServerHandler<rpc::RPCEngine, etl::ETLService>>(
+        config, backend, rpcEngine, etl, subscriptions);
     auto ctx = parseCerts(config);
     auto const ctxRef = ctx ? std::optional<std::reference_wrapper<ssl::context>>{ctx.value()} : std::nullopt;
-    auto const httpServer = Server::make_HttpServer(config, ioc, ctxRef, dosGuard, executor);
+    auto const httpServer = web::make_HttpServer(config, ioc, ctxRef, dosGuard, handler);
 
     // Blocks until stopped.
     // When stopped, shared_ptrs fall out of scope
@@ -221,5 +224,5 @@ try
 }
 catch (std::exception const& e)
 {
-    LogService::fatal() << "Exit on exception: " << e.what();
+    LOG(LogService::fatal()) << "Exit on exception: " << e.what();
 }

src/main/Mainpage.h

@@ -0,0 +1,43 @@
+//------------------------------------------------------------------------------
+/*
+    This file is part of clio: https://github.com/XRPLF/clio
+    Copyright (c) 2023, the clio developers.
+
+    Permission to use, copy, modify, and distribute this software for any
+    purpose with or without fee is hereby granted, provided that the above
+    copyright notice and this permission notice appear in all copies.
+
+    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
+    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
+    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+//==============================================================================
+
+/**
+ * @mainpage Clio API server
+ *
+ * @section intro Introduction
+ *
+ * Clio is an XRP Ledger API server. Clio is optimized for RPC calls, over WebSocket or JSON-RPC.
+ *
+ * Validated historical ledger and transaction data are stored in a more space-efficient format, using up to 4 times
+ * less space than rippled.
+ *
+ * Clio can be configured to store data in Apache Cassandra or ScyllaDB, allowing for scalable read throughput.
+ * Multiple Clio nodes can share access to the same dataset, allowing for a highly available cluster of Clio nodes,
+ * without the need for redundant data storage or computation.
+ *
+ * You can read more general information about Clio and its subsystems from the `Related Pages` section.
+ *
+ * @section Develop
+ *
+ * As you prepare to develop code for Clio, please be sure you are aware of our current
+ * <A HREF="https://github.com/XRPLF/clio/blob/develop/CONTRIBUTING.md">Contribution guidelines</A>.
+ *
+ * Read `rpc/README.md` carefully to know more about writing your own handlers for
+ * Clio.
+ */

src/rpc/Amendments.h

@@ -0,0 +1,48 @@
+//------------------------------------------------------------------------------
+/*
+    This file is part of clio: https://github.com/XRPLF/clio
+    Copyright (c) 2023, the clio developers.
+
+    Permission to use, copy, modify, and distribute this software for any
+    purpose with or without fee is hereby granted, provided that the above
+    copyright notice and this permission notice appear in all copies.
+
+    THE  SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+    WITH  REGARD  TO  THIS  SOFTWARE  INCLUDING  ALL  IMPLIED  WARRANTIES  OF
+    MERCHANTABILITY  AND  FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+    ANY  SPECIAL,  DIRECT,  INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+    WHATSOEVER  RESULTING  FROM  LOSS  OF USE, DATA OR PROFITS, WHETHER IN AN
+    ACTION  OF  CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+    OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+//==============================================================================
+
+#pragma once
+
+#include <ripple/basics/base_uint.h>
+
+#include <ripple/protocol/digest.h>
+
+namespace rpc {
+
+#define REGISTER_AMENDMENT(name) inline static const ripple::uint256 name = GetAmendmentId(#name);
+
+/**
+ * @brief Represents a list of amendments in the XRPL.
+ */
+struct Amendments
+{
+    /**
+     * @param name The name of the amendment
+     * @return The corresponding amendment Id
+     */
+    static ripple::uint256 const
+    GetAmendmentId(std::string_view const name)
+    {
+        return ripple::sha512Half(ripple::Slice(name.data(), name.size()));
+    }
+
+    REGISTER_AMENDMENT(DisallowIncoming)
+    REGISTER_AMENDMENT(Clawback)
+};
+}  // namespace rpc

src/rpc/BookChangesHelper.h

@@ -17,13 +17,14 @@
 */
 //==============================================================================
 
+/** @file */
 #pragma once
 
 #include <rpc/RPCHelpers.h>
 
 #include <set>
 
-namespace RPC {
+namespace rpc {
 
 /**
  * @brief Represents an entry in the book_changes' changes array.
@@ -53,7 +54,7 @@ public:
      * @return std::vector<BookChange> Book changes
      */
     [[nodiscard]] static std::vector<BookChange>
-    compute(std::vector<Backend::TransactionAndMetadata> const& transactions)
+    compute(std::vector<data::TransactionAndMetadata> const& transactions)
     {
         return HandlerImpl{}(transactions);
     }
@@ -66,7 +67,7 @@ private:
 
     public:
         [[nodiscard]] std::vector<BookChange>
-        operator()(std::vector<Backend::TransactionAndMetadata> const& transactions)
+        operator()(std::vector<data::TransactionAndMetadata> const& transactions)
         {
             for (auto const& tx : transactions)
                 handleBookChange(tx);
@@ -164,22 +165,21 @@ private:
             }
             else
             {
-                // TODO: use paranthesized initialization when clang catches up
                 tally_[key] = {
-                    first,   // sideAVolume
-                    second,  // sideBVolume
-                    rate,    // highRate
-                    rate,    // lowRate
-                    rate,    // openRate
-                    rate,    // closeRate
+                    .sideAVolume = first,
+                    .sideBVolume = second,
+                    .highRate = rate,
+                    .lowRate = rate,
+                    .openRate = rate,
+                    .closeRate = rate,
                 };
             }
         }
 
         void
-        handleBookChange(Backend::TransactionAndMetadata const& blob)
+        handleBookChange(data::TransactionAndMetadata const& blob)
         {
-            auto const [tx, meta] = RPC::deserializeTxPlusMeta(blob);
+            auto const [tx, meta] = rpc::deserializeTxPlusMeta(blob);
             if (!tx || !meta || !tx->isFieldPresent(ripple::sfTransactionType))
                 return;
 
@@ -199,6 +199,7 @@ private:
                 case ripple::ttOFFER_CREATE:
                     if (tx->isFieldPresent(ripple::sfOfferSequence))
                         return tx->getFieldU32(ripple::sfOfferSequence);
+                    [[fallthrough]];
                 default:
                     return std::nullopt;
             }
@@ -206,6 +207,12 @@ private:
     };
 };
 
+/**
+ * @brief Implementation of value_from for BookChange type.
+ *
+ * @param jv The JSON value to populate
+ * @param change The BookChange to serialize
+ */
 inline void
 tag_invoke(boost::json::value_from_tag, boost::json::value& jv, BookChange const& change)
 {
@@ -229,7 +236,13 @@ tag_invoke(boost::json::value_from_tag, boost::json::value& jv, BookChange const
     };
 }
 
+/**
+ * @brief Computes all book changes for the given ledger header and transactions.
+ *
+ * @param lgrInfo The ledger header
+ * @param transactions The vector of transactions with heir metadata
+ */
 [[nodiscard]] boost::json::object const
-computeBookChanges(ripple::LedgerInfo const& lgrInfo, std::vector<Backend::TransactionAndMetadata> const& transactions);
+computeBookChanges(ripple::LedgerHeader const& lgrInfo, std::vector<data::TransactionAndMetadata> const& transactions);
 
-}  // namespace RPC
+}  // namespace rpc