Update some of the invariant checks to account for leaves

.gitea/workflows/ci.yml

@@ -1,339 +0,0 @@
-name: CI
-
-on: [push, pull_request]
-
-env:
-  CC: clang
-  CXX: clang++
-
-jobs:
-  pre-commit:
-    runs-on: ubuntu-latest-amd64
-    steps:
-      - uses: actions/checkout@v4
-
-      - uses: actions/cache@v4
-        with:
-          path: /var/cache/apt/archives
-          key: apt-amd64-${{ hashFiles('.gitea/workflows/ci.yml') }}
-
-      - name: Install dependencies
-        run: |
-          sudo apt-get update -qq
-          sudo apt-get install -y git nodejs pre-commit
-
-      - uses: actions/cache@v4
-        with:
-          path: ~/.cache/pre-commit
-          key: pre-commit-${{ hashFiles('.pre-commit-config.yaml') }}
-
-      - name: Run pre-commit
-        run: |
-          git config --global --add safe.directory "$PWD"
-          pre-commit run --all-files --show-diff-on-failure
-
-  test:
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - name: 64-bit-versions
-            runner: ubuntu-latest-amd64
-            arch: amd64
-            llvm_version: "21"
-            msan_url: ""
-            cmake_args: -DCMAKE_CXX_FLAGS=-DUSE_64_BIT=1
-          - name: debug
-            runner: ubuntu-latest-amd64
-            arch: amd64
-            llvm_version: "21"
-            msan_url: https://minio.weaselab.dev/public/x86_64/msan-toolchain-21.1.8.tar.zst
-            cmake_args: -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan
-          - name: simd-fallback
-            runner: ubuntu-latest-amd64
-            arch: amd64
-            llvm_version: "21"
-            msan_url: ""
-            cmake_args: -DUSE_SIMD_FALLBACK=ON
-          - name: gcc
-            runner: ubuntu-latest-amd64
-            arch: amd64
-            llvm_version: "21"
-            msan_url: ""
-            cmake_args: -DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++
-          - name: debug-arm64
-            runner: ubuntu-latest-arm64
-            arch: arm64
-            llvm_version: "22"
-            msan_url: https://minio.weaselab.dev/public/aarch64/msan-toolchain-22.1.8.tar.zst
-            cmake_args: -DCMAKE_BUILD_TYPE=Debug -DMSAN_TOOLCHAIN_PATH=/opt/msan
-    runs-on: ${{ matrix.runner }}
-    steps:
-      - uses: actions/checkout@v4
-
-      - uses: actions/cache@v4
-        with:
-          path: /var/cache/apt/archives
-          key: apt-${{ matrix.arch }}-${{ hashFiles('.gitea/workflows/ci.yml') }}
-
-      - name: Install common dependencies
-        run: |
-          sudo apt-get update -qq
-          sudo apt-get install -y \
-            build-essential ccache cmake gcc g++ libc6-dbg \
-            ninja-build python3 valgrind zstd
-          sudo curl -Ls "https://minio.weaselab.dev/public/$(uname -m)/mc.RELEASE.2025-08-13T08-35-41Z" \
-            -o /usr/local/bin/mc && sudo chmod +x /usr/local/bin/mc
-
-      - name: Install LLVM toolchain
-        if: matrix.name != 'gcc'
-        run: |
-          . /etc/os-release
-          wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key | sudo tee /etc/apt/trusted.gpg.d/apt.llvm.org.asc
-          echo "deb http://apt.llvm.org/${VERSION_CODENAME}/ llvm-toolchain-${VERSION_CODENAME}-${{ matrix.llvm_version }} main" | sudo tee /etc/apt/sources.list.d/llvm.list
-          sudo apt-get update -qq
-          sudo apt-get install -y \
-            clang-${{ matrix.llvm_version }} llvm-${{ matrix.llvm_version }} lld-${{ matrix.llvm_version }} mold
-          for tool in clang clang++ llvm-ar llvm-nm llvm-ranlib llvm-objcopy llvm-cov llvm-symbolizer lld ld.lld; do
-            sudo update-alternatives --install /usr/bin/${tool} ${tool} /usr/bin/${tool}-${{ matrix.llvm_version }} 100
-          done
-
-      - name: Download MSan toolchain
-        if: matrix.msan_url != ''
-        run: |
-          curl -Ls "${{ matrix.msan_url }}" -o /tmp/msan-toolchain.tar.zst
-          sudo mkdir -p /opt/msan
-          sudo tar --zstd -xf /tmp/msan-toolchain.tar.zst -C /opt/msan
-
-      - name: Build
-        run: |
-          export CCACHE_DIR="$GITHUB_WORKSPACE/.ccache"
-          rm -rf build
-          cmake -S . -B build -G Ninja -DCMAKE_CXX_COMPILER_LAUNCHER=ccache ${{ matrix.cmake_args }}
-          ninja -C build
-          ccache -s
-
-      - name: Test
-        run: |
-          cd build
-          ctest --no-compress-output --test-output-size-passed 100000 --test-output-size-failed 100000 -T Test -j "$(nproc)" --timeout 90 > /dev/null
-
-      - name: Upload test results to MinIO
-        if: always()
-        env:
-          MINIO_ACCESS_KEY: ${{ secrets.MINIO_ACCESS_KEY }}
-          MC_HOST_minio: https://${{ secrets.MINIO_ACCESS_KEY }}:${{ secrets.MINIO_SECRET_KEY }}@minio.weaselab.dev
-        run: |
-          if [ -z "$MINIO_ACCESS_KEY" ]; then
-            echo "MinIO credentials not configured; skipping upload"
-            exit 0
-          fi
-          zstd build/Testing/*/Test.xml
-          mc cp build/Testing/*/Test.xml.zst "minio/jenkins/conflict-set/${{ gitea.run_number }}/${{ matrix.name }}/"
-
-      - name: Test summary
-        if: always()
-        run: |
-          python3 ctest_summary.py build/Testing/*/Test.xml \
-            --link "https://minio.weaselab.dev/jenkins/conflict-set/${{ gitea.run_number }}/${{ matrix.name }}/Test.xml.zst" \
-            | tee -a "$GITHUB_STEP_SUMMARY"
-
-  release:
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - runner: ubuntu-latest-amd64
-            arch: amd64
-            cmake_args: -DMSAN_TOOLCHAIN_PATH=/opt/msan
-          - runner: ubuntu-latest-arm64
-            arch: arm64
-            cmake_args: ""
-    runs-on: ${{ matrix.runner }}
-    steps:
-      - uses: actions/checkout@v4
-
-      - uses: actions/cache@v4
-        with:
-          path: /var/cache/apt/archives
-          key: apt-${{ matrix.arch }}-${{ hashFiles('.gitea/workflows/ci.yml') }}
-
-      - name: Install dependencies
-        run: |
-          . /etc/os-release
-          wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key | sudo tee /etc/apt/trusted.gpg.d/apt.llvm.org.asc
-          echo "deb http://apt.llvm.org/${VERSION_CODENAME}/ llvm-toolchain-${VERSION_CODENAME}-21 main" | sudo tee /etc/apt/sources.list.d/llvm.list
-          sudo apt-get update -qq
-          sudo apt-get install -y \
-            build-essential ccache clang-21 cmake devscripts \
-            libc6-dbg llvm-21 lld-21 mold ninja-build rpm \
-            valgrind zstd
-          sudo curl -Ls "https://minio.weaselab.dev/public/$(uname -m)/mc.RELEASE.2025-08-13T08-35-41Z" \
-            -o /usr/local/bin/mc && sudo chmod +x /usr/local/bin/mc
-          for tool in clang clang++ llvm-ar llvm-nm llvm-ranlib llvm-objcopy llvm-cov llvm-symbolizer lld ld.lld; do
-            sudo update-alternatives --install /usr/bin/${tool} ${tool} /usr/bin/${tool}-21 100
-          done
-
-      - name: Install paper build dependencies
-        if: matrix.arch == 'amd64'
-        run: |
-          sudo apt-get install -y \
-            biber latexmk \
-            texlive-bibtex-extra texlive-fonts-recommended \
-            texlive-latex-extra texlive-pictures
-
-      - uses: actions/cache@v4
-        with:
-          path: .ccache
-          key: ccache-release-${{ matrix.arch }}-${{ gitea.sha }}
-          restore-keys: |
-            ccache-release-${{ matrix.arch }}-
-
-      - name: Download MSan toolchain
-        if: matrix.arch == 'amd64'
-        run: |
-          curl -Ls "https://minio.weaselab.dev/public/x86_64/msan-toolchain-21.1.8.tar.zst" -o /tmp/msan-toolchain.tar.zst
-          sudo mkdir -p /opt/msan
-          sudo tar --zstd -xf /tmp/msan-toolchain.tar.zst -C /opt/msan
-
-      - name: Build
-        run: |
-          export CCACHE_DIR="$GITHUB_WORKSPACE/.ccache"
-          rm -rf build
-          cmake -S . -B build -G Ninja -DCMAKE_CXX_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_FLAGS=-DNVALGRIND ${{ matrix.cmake_args }}
-          ninja -C build
-          ccache -s
-
-      - name: Test
-        run: |
-          cd build
-          ctest --no-compress-output --test-output-size-passed 100000 --test-output-size-failed 100000 -T Test -j "$(nproc)" --timeout 90 > /dev/null
-      - name: Package
-        run: |
-          cd build
-          cpack -G DEB
-          cpack -G RPM
-
-      - name: Build paper
-        if: matrix.arch == 'amd64'
-        run: |
-          cd paper
-          make
-
-      - name: Upload artifacts to MinIO
-        if: always()
-        shell: bash
-        env:
-          MINIO_ACCESS_KEY: ${{ secrets.MINIO_ACCESS_KEY }}
-          MC_HOST_minio: https://${{ secrets.MINIO_ACCESS_KEY }}:${{ secrets.MINIO_SECRET_KEY }}@minio.weaselab.dev
-        run: |
-          if [ -z "$MINIO_ACCESS_KEY" ]; then
-            echo "MinIO credentials not configured; skipping upload"
-            exit 0
-          fi
-          dest="minio/jenkins/conflict-set/${{ gitea.run_number }}/release-${{ matrix.arch }}/"
-          zstd build/Testing/*/Test.xml
-          mc cp build/Testing/*/Test.xml.zst "$dest"
-          if compgen -G "build/*.deb" > /dev/null; then
-            mc cp build/*.deb "$dest"
-          fi
-          if compgen -G "build/*.rpm" > /dev/null; then
-            mc cp build/*.rpm "$dest"
-          fi
-          if compgen -G "paper/*.pdf" > /dev/null; then
-            mc cp paper/*.pdf "$dest"
-          fi
-
-      - name: Test summary
-        if: always()
-        run: |
-          python3 ctest_summary.py build/Testing/*/Test.xml \
-            --link "https://minio.weaselab.dev/jenkins/conflict-set/${{ gitea.run_number }}/release-${{ matrix.arch }}/Test.xml.zst" \
-            | tee -a "$GITHUB_STEP_SUMMARY"
-
-  coverage:
-    runs-on: ubuntu-latest-amd64
-    steps:
-      - uses: actions/checkout@v4
-
-      - uses: actions/cache@v4
-        with:
-          path: /var/cache/apt/archives
-          key: apt-amd64-${{ hashFiles('.gitea/workflows/ci.yml') }}
-
-      - name: Install dependencies
-        run: |
-          . /etc/os-release
-          wget -qO- https://apt.llvm.org/llvm-snapshot.gpg.key | sudo tee /etc/apt/trusted.gpg.d/apt.llvm.org.asc
-          echo "deb http://apt.llvm.org/${VERSION_CODENAME}/ llvm-toolchain-${VERSION_CODENAME}-21 main" | sudo tee /etc/apt/sources.list.d/llvm.list
-          sudo apt-get update -qq
-          sudo apt-get install -y \
-            build-essential ccache clang-21 cmake gcovr \
-            libc6-dbg llvm-21 lld-21 mold ninja-build python3 valgrind zstd
-          sudo curl -Ls "https://minio.weaselab.dev/public/$(uname -m)/mc.RELEASE.2025-08-13T08-35-41Z" \
-            -o /usr/local/bin/mc && sudo chmod +x /usr/local/bin/mc
-          for tool in clang clang++ llvm-ar llvm-nm llvm-ranlib llvm-objcopy llvm-cov llvm-symbolizer lld ld.lld; do
-            sudo update-alternatives --install /usr/bin/${tool} ${tool} /usr/bin/${tool}-21 100
-          done
-
-      - uses: actions/cache@v4
-        with:
-          path: .ccache
-          key: ccache-coverage-${{ gitea.sha }}
-          restore-keys: |
-            ccache-coverage-
-
-      - name: Build
-        run: |
-          export CCACHE_DIR="$GITHUB_WORKSPACE/.ccache"
-          rm -rf build
-          cmake -S . -B build -G Ninja -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
-            -DCMAKE_C_FLAGS=--coverage -DCMAKE_CXX_FLAGS=--coverage \
-            -DCMAKE_BUILD_TYPE=Debug -DDISABLE_TSAN=ON
-          ninja -C build
-          ccache -s
-
-      - name: Test
-        run: |
-          cd build
-          ctest --no-compress-output --test-output-size-passed 100000 --test-output-size-failed 100000 -T Test -j "$(nproc)" --timeout 90 > /dev/null
-
-      - name: Coverage report
-        shell: bash
-        run: |
-          gcov_args=(-f ConflictSet.cpp -f LongestCommonPrefix.h -f Metrics.h
-                     --gcov-executable "llvm-cov gcov" --exclude-noncode-lines)
-          gcovr "${gcov_args[@]}" --cobertura > build/coverage.xml
-          gcovr "${gcov_args[@]}"
-          mkdir -p build/coverage_html
-          gcovr "${gcov_args[@]}" --html-details build/coverage_html/index.html
-          gcovr "${gcov_args[@]}" --fail-under-line 100 > /dev/null
-
-      - name: Upload results to MinIO
-        if: always()
-        env:
-          MINIO_ACCESS_KEY: ${{ secrets.MINIO_ACCESS_KEY }}
-          MC_HOST_minio: https://${{ secrets.MINIO_ACCESS_KEY }}:${{ secrets.MINIO_SECRET_KEY }}@minio.weaselab.dev
-        run: |
-          if [ -z "$MINIO_ACCESS_KEY" ]; then
-            echo "MinIO credentials not configured; skipping upload"
-            exit 0
-          fi
-          dest="minio/jenkins/conflict-set/${{ gitea.run_number }}/coverage/"
-          zstd build/Testing/*/Test.xml
-          mc cp build/Testing/*/Test.xml.zst "$dest"
-          if [ -e build/coverage.xml ]; then
-            mc cp build/coverage.xml "$dest"
-          fi
-          if [ -d build/coverage_html ]; then
-            mc cp -r build/coverage_html "$dest"
-          fi
-
-      - name: Test summary
-        if: always()
-        run: |
-          python3 ctest_summary.py build/Testing/*/Test.xml \
-            --link "https://minio.weaselab.dev/jenkins/conflict-set/${{ gitea.run_number }}/coverage/Test.xml.zst" \
-            | tee -a "$GITHUB_STEP_SUMMARY"
-          echo "" | tee -a "$GITHUB_STEP_SUMMARY"
-          echo "📊 [Coverage report](https://minio.weaselab.dev/jenkins/conflict-set/${{ gitea.run_number }}/coverage/coverage_html/index.html)" | tee -a "$GITHUB_STEP_SUMMARY"

Bench.cpp

@@ -37,16 +37,15 @@ ConflictSet::ReadRange singleton(Arena &arena, TrivialSpan key) {
 }
 
 ConflictSet::ReadRange prefixRange(Arena &arena, TrivialSpan key) {
-  int index = key.size() - 1;
+  int index;
-  for (; index >= 0; index--)
+  for (index = key.size() - 1; index >= 0; index--)
-    if (key[index] != 255)
+    if ((key[index]) != 255)
       break;
 
   // Must not be called with a string that consists only of zero or more '\xff'
-  // bytes, or with an empty string (which has no finite upper bound).
+  // bytes.
   if (index < 0) {
     assert(false);
-    std::abort();
   }
 
   uint8_t *buf = new (arena) uint8_t[index + 1];

CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.18)
 project(
   conflict-set
-  VERSION 0.1.0
+  VERSION 0.0.14
   DESCRIPTION
     "A data structure for optimistic concurrency control on ranges of bitwise-lexicographically-ordered keys."
   HOMEPAGE_URL "https://git.weaselab.dev/weaselab/conflict-set"
@@ -32,12 +32,8 @@ if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
 endif()
 
 add_compile_options(
-  -Werror=switch-enum
+  # -Werror=switch-enum
-  -Wswitch-enum
+  -Wswitch-enum -Wunused-variable -fPIC -fdata-sections -ffunction-sections
-  -Wunused-variable
-  -fPIC
-  -fdata-sections
-  -ffunction-sections
   -fno-jump-tables # https://github.com/llvm/llvm-project/issues/54247
 )
 
@@ -53,7 +49,6 @@ endif()
 
 if(CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
   add_compile_options("-Wno-maybe-uninitialized")
-  add_compile_options("-Wno-maybe-musttail-local-addr")
 endif()
 
 if(NOT APPLE)
@@ -61,16 +56,12 @@ if(NOT APPLE)
   add_compile_options(-g -fno-omit-frame-pointer)
 endif()
 
-set(relro_flags "LINKER:-z,relro,-z,now,-z,noexecstack")
+set(full_relro_flags "-pie;LINKER:-z,relro,-z,now,-z,noexecstack")
-set(full_relro_flags "-pie;${relro_flags}")
 cmake_push_check_state()
 list(APPEND CMAKE_REQUIRED_LINK_OPTIONS ${full_relro_flags})
 check_cxx_source_compiles("int main(){}" HAS_FULL_RELRO FAIL_REGEX "warning:")
 if(HAS_FULL_RELRO)
-  # -pie only applies to executables; passing it when linking a shared library
+  add_link_options(${full_relro_flags})
-  # makes the driver pull in Scrt1.o, which requires main.
-  add_link_options("$<$<STREQUAL:$<TARGET_PROPERTY:TYPE>,EXECUTABLE>:-pie>"
-                   ${relro_flags})
 endif()
 cmake_pop_check_state()
 
@@ -103,13 +94,6 @@ option(USE_SIMD_FALLBACK
 
 option(DISABLE_TSAN "Disable TSAN" OFF)
 
-set(MSAN_TOOLCHAIN_PATH
-    ""
-    CACHE
-      PATH
-      "Path to an MSan-instrumented libc++ toolchain. When set, an MSan fuzz_driver target is added. Must contain include/c++/v1 and lib. See build_msan_toolchain.sh to produce one."
-)
-
 # This is encouraged according to
 # https://valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.clientreq
 include_directories(SYSTEM ${CMAKE_CURRENT_SOURCE_DIR}/third_party/valgrind)
@@ -146,7 +130,6 @@ add_custom_command(
   COMMAND_EXPAND_LISTS)
 
 add_library(${PROJECT_NAME} SHARED ${CMAKE_BINARY_DIR}/${PROJECT_NAME}.o)
-add_dependencies(${PROJECT_NAME} ${PROJECT_NAME}-object)
 set_target_properties(
   ${PROJECT_NAME} PROPERTIES LIBRARY_OUTPUT_DIRECTORY
                              "${CMAKE_CURRENT_BINARY_DIR}/radix_tree")
@@ -163,7 +146,6 @@ if(HAS_VERSION_SCRIPT)
 endif()
 
 add_library(${PROJECT_NAME}-static STATIC ${CMAKE_BINARY_DIR}/${PROJECT_NAME}.o)
-add_dependencies(${PROJECT_NAME}-static ${PROJECT_NAME}-object)
 if(CMAKE_BUILD_TYPE STREQUAL Debug)
   set_target_properties(${PROJECT_NAME}-static PROPERTIES LINKER_LANGUAGE CXX)
 else()
@@ -182,8 +164,6 @@ endif()
 
 include(CTest)
 
-option(TEST_SKIP_LIST "Run the slow skip list cross-check tests" OFF)
-
 # disable tests if this is being used through e.g. FetchContent
 if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR AND BUILD_TESTING)
 
@@ -226,9 +206,8 @@ if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR AND BUILD_TESTING)
   target_compile_options(driver_skip_list PRIVATE ${TEST_FLAGS})
   target_link_libraries(driver_skip_list PRIVATE skip_list)
 
-  # Optional cross-check of SkipList.cpp against ReferenceImpl. Opt-in via
+  # enable to test skip list
-  # -DTEST_SKIP_LIST=ON, since these are slow.
+  if(0)
-  if(TEST_SKIP_LIST)
     foreach(TEST ${CORPUS_TESTS})
       get_filename_component(hash ${TEST} NAME)
       add_test(NAME skip_list_${hash} COMMAND driver_skip_list ${TEST})
@@ -263,7 +242,7 @@ if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR AND BUILD_TESTING)
     endif()
   endif()
 
-  # whitebox tests asan+ubsan
+  # whitebox tests
   add_executable(fuzz_driver ConflictSet.cpp FuzzTestDriver.cpp)
   target_compile_options(fuzz_driver PRIVATE ${TEST_FLAGS})
   if(NOT CMAKE_CROSSCOMPILING)
@@ -278,34 +257,6 @@ if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR AND BUILD_TESTING)
     add_test(NAME conflict_set_fuzz_${hash} COMMAND fuzz_driver ${TEST})
   endforeach()
 
-  # whitebox tests msan
-  if(MSAN_TOOLCHAIN_PATH)
-    add_executable(fuzz_driver_msan ConflictSet.cpp FuzzTestDriver.cpp)
-    target_compile_options(fuzz_driver_msan PRIVATE ${TEST_FLAGS})
-    if(NOT CMAKE_CROSSCOMPILING)
-      target_compile_options(
-        fuzz_driver_msan
-        PRIVATE -fsanitize=memory -fsanitize-memory-track-origins=2
-                $<$<COMPILE_LANGUAGE:CXX>:-stdlib=libc++>
-                -I${MSAN_TOOLCHAIN_PATH}/include/c++/v1)
-      target_link_options(
-        fuzz_driver_msan
-        PRIVATE
-        -fsanitize=memory
-        -fsanitize-memory-track-origins=2
-        -stdlib=libc++
-        -L${MSAN_TOOLCHAIN_PATH}/lib
-        LINKER:-rpath,${MSAN_TOOLCHAIN_PATH}/lib)
-    endif()
-    target_compile_definitions(fuzz_driver_msan PRIVATE ENABLE_FUZZ)
-    target_include_directories(fuzz_driver_msan
-                               PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
-    foreach(TEST ${CORPUS_TESTS})
-      get_filename_component(hash ${TEST} NAME)
-      add_test(NAME conflict_set_msan_${hash} COMMAND fuzz_driver_msan ${TEST})
-    endforeach()
-  endif()
-
   # tsan tests
   if(NOT CMAKE_CROSSCOMPILING AND NOT DISABLE_TSAN)
     add_executable(tsan_driver ConflictSet.cpp FuzzTestDriver.cpp)
@@ -393,24 +344,6 @@ if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR AND BUILD_TESTING)
                         PROPERTIES CXX_STANDARD_REQUIRED ON)
   add_test(NAME conflict_set_cxx_api_test COMMAND conflict_set_cxx_api_test)
 
-  # SHOW_MEMORY instrumentation, which the rest of the test suite doesn't
-  # compile with
-  add_executable(conflict_set_show_memory_test ConflictSet.cpp
-                                               ShowMemoryTest.cpp)
-  target_compile_options(conflict_set_show_memory_test PRIVATE ${TEST_FLAGS})
-  target_compile_definitions(conflict_set_show_memory_test
-                             PRIVATE SHOW_MEMORY=1)
-  target_include_directories(conflict_set_show_memory_test
-                             PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
-  if(NOT CMAKE_CROSSCOMPILING)
-    target_compile_options(conflict_set_show_memory_test
-                           PRIVATE -fsanitize=address,undefined)
-    target_link_options(conflict_set_show_memory_test PRIVATE
-                        -fsanitize=address,undefined)
-  endif()
-  add_test(NAME conflict_set_show_memory_test
-           COMMAND conflict_set_show_memory_test)
-
   # symbol visibility tests
   if(NOT CMAKE_BUILD_TYPE STREQUAL Debug)
     if(APPLE)
@@ -441,30 +374,9 @@ if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR AND BUILD_TESTING)
   if(NOT CMAKE_CROSSCOMPILING)
     find_program(HARDENING_CHECK hardening-check)
     if(HARDENING_CHECK)
-      # Not all versions of hardening-check support the same options, so query
+      add_test(NAME hardening_check
-      # the help output before using architecture-specific skips.
+               COMMAND ${HARDENING_CHECK} $<TARGET_FILE:${PROJECT_NAME}>
-      execute_process(
+                       --nofortify --nostackprotector)
-        COMMAND ${HARDENING_CHECK} --help
-        OUTPUT_VARIABLE hardening_check_help
-        ERROR_VARIABLE hardening_check_help
-        OUTPUT_STRIP_TRAILING_WHITESPACE ERROR_STRIP_TRAILING_WHITESPACE)
-      set(hardening_check_arch_flags "")
-      # Control flow integrity (CET) is x86-only and branch protection (PAC/BTI)
-      # is arm64-only, so ignore whichever doesn't apply.
-      if(CMAKE_SYSTEM_PROCESSOR STREQUAL aarch64 OR CMAKE_SYSTEM_PROCESSOR
-                                                    STREQUAL arm64)
-        if(hardening_check_help MATCHES "nocfprotection")
-          list(APPEND hardening_check_arch_flags --nocfprotection)
-        endif()
-      else()
-        if(hardening_check_help MATCHES "nobranchprotection")
-          list(APPEND hardening_check_arch_flags --nobranchprotection)
-        endif()
-      endif()
-      add_test(
-        NAME hardening_check
-        COMMAND ${HARDENING_CHECK} $<TARGET_FILE:${PROJECT_NAME}> --nofortify
-                --nostackprotector ${hardening_check_arch_flags})
     endif()
   endif()
 

Dockerfile

@@ -0,0 +1,58 @@
+FROM ubuntu:rolling
+
+WORKDIR /tmp
+ENV HOME=/tmp
+RUN chmod -R 777 /tmp
+
+# Install apt dependencies
+RUN apt-get update
+RUN apt-get upgrade -y
+RUN TZ=America/Los_Angeles DEBIAN_FRONTEND=noninteractive apt-get install -y \
+    binutils-aarch64-linux-gnu \
+    build-essential \
+    ccache \
+    cmake \
+    curl \
+    devscripts \
+    g++-aarch64-linux-gnu \
+    gcovr \
+    git \
+    gnupg \
+    libc6-dbg \
+    lsb-release \
+    mold \
+    ninja-build \
+    pre-commit \
+    python3-requests \
+    qemu-user \
+    rpm \
+    software-properties-common \
+    texlive-full \
+    wget \
+    zstd
+
+# Install recent valgrind from source
+RUN curl -Ls https://sourceware.org/pub/valgrind/valgrind-3.22.0.tar.bz2 -o valgrind.tar.bz2 && \
+    echo "c811db5add2c5f729944caf47c4e7a65dcaabb9461e472b578765dd7bf6d2d4c  valgrind.tar.bz2" > valgrind-sha.txt && \
+    sha256sum --quiet -c valgrind-sha.txt && \
+    mkdir valgrind && \
+    tar --strip-components 1 --no-same-owner --no-same-permissions --directory valgrind -xjf valgrind.tar.bz2 && \
+    cd valgrind && \
+    ./configure --enable-only64bit --enable-lto && \
+    make -j`nproc` && \
+    make install && \
+    cd .. && \
+    rm -rf /tmp/*
+
+# Recent clang
+RUN wget https://apt.llvm.org/llvm.sh && chmod +x ./llvm.sh && ./llvm.sh 20
+
+RUN apt-get -y install clang llvm
+
+# Set after building valgrind, which doesn't build with clang for some reason
+ENV CC=clang
+ENV CXX=clang++
+
+# Try to have all the pre-commit hooks we'll need already initialized
+COPY .pre-commit-config.yaml /tmp/
+RUN git init && pre-commit install-hooks

HashTable.cpp

@@ -59,12 +59,6 @@ struct __attribute__((visibility("hidden"))) ConflictSet::Impl {
     while (keyUpdates > 0) {
       if (iter == map.end()) {
         iter = map.begin();
-        // The map is empty, so there are no more entries to garbage-collect.
-        // Break rather than spinning forever (keyUpdates is charged at 2x
-        // the entry count, so it may still be > 0 after emptying the map).
-        if (iter == map.end()) {
-          break;
-        }
       }
       for (; iter != map.end(); --keyUpdates) {
         if (iter->second <= oldestVersion) {
@@ -102,9 +96,7 @@ void ConflictSet::setOldestVersion(int64_t oldestVersion) {
   return impl->setOldestVersion(oldestVersion);
 }
 
-// The hash_table implementation does not track memory usage, so return 0 to
+int64_t ConflictSet::getBytes() const { return -1; }
-// satisfy the API contract that getBytes() returns a non-negative value.
-int64_t ConflictSet::getBytes() const { return 0; }
 
 void ConflictSet::getMetricsV1(MetricsV1 **metrics, int *count) const {
   *metrics = nullptr;
@@ -127,13 +119,7 @@ ConflictSet::ConflictSet(ConflictSet &&other) noexcept
     : impl(std::exchange(other.impl, nullptr)) {}
 
 ConflictSet &ConflictSet::operator=(ConflictSet &&other) noexcept {
-  if (this != &other) {
-    if (impl) {
-      impl->~Impl();
-      safe_free(impl, sizeof(Impl));
-    }
   impl = std::exchange(other.impl, nullptr);
-  }
   return *this;
 }
 
@@ -169,10 +155,7 @@ __attribute__((__visibility__("default"))) void ConflictSet_destroy(void *cs) {
 }
 __attribute__((__visibility__("default"))) int64_t
 ConflictSet_getBytes(void *cs) {
-  (void)cs;
+  using Impl = ConflictSet::Impl;
-  // The hash_table implementation does not track memory usage, so return 0 to
+  return -1;
-  // satisfy the API contract that ConflictSet_getBytes returns a non-negative
-  // value.
-  return 0;
 }
 }

Internal.h

@@ -23,9 +23,7 @@ using namespace weaselab;
 #include <callgrind.h>
 
 #define DEBUG_VERBOSE 0
-#ifndef SHOW_MEMORY
 #define SHOW_MEMORY 0
-#endif
 
 // std::span is not trivially constructible. We want a span that leaves its
 // members uninitialized for performance reasons.
@@ -98,11 +96,6 @@ static_assert(std::is_trivial_v<TrivialSpan>);
 #if SHOW_MEMORY
 inline int64_t mallocBytes = 0;
 inline int64_t peakMallocBytes = 0;
-// Defined in ConflictSet.cpp
-extern int64_t nodeBytes;
-extern int64_t partialCapacityBytes;
-extern int64_t totalKeys;
-extern int64_t keyBytes;
 #endif
 
 inline thread_local int64_t mallocBytesDelta = 0;

Jenkinsfile

@@ -0,0 +1,151 @@
+def CleanBuildAndTest(String cmakeArgs) {
+    sh """
+    export CCACHE_DIR=/ccache
+    rm -rf build
+    mkdir build
+    cd build
+    cmake .. -G Ninja -DCMAKE_CXX_COMPILER_LAUNCHER=ccache ${cmakeArgs} 
+    ninja
+    ccache -s
+    """
+    catchError {
+        sh '''
+        cd build
+        ctest --no-compress-output --test-output-size-passed 100000 --test-output-size-failed 100000 -T Test -j `nproc` --timeout 90 > /dev/null
+        zstd Testing/*/Test.xml
+        '''
+    }
+    xunit tools: [CTest(pattern: 'build/Testing/*/Test.xml')], skipPublishingChecks: false
+    minio bucket: 'jenkins', credentialsId: 'jenkins-minio', excludes: '', host: 'minio.weaselab.dev', includes: 'build/Testing/*/Test.xml.zst', targetFolder: '${JOB_NAME}/${BUILD_NUMBER}/${STAGE_NAME}/'
+}
+
+pipeline {
+    agent any
+    stages {
+        stage('Pre-commit') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                script {
+                    env.HOME = env.WORKSPACE
+                }
+                sh 'pre-commit run --all-files --show-diff-on-failure'
+            }
+        }
+        stage('64 bit versions') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                CleanBuildAndTest("-DCMAKE_CXX_FLAGS=-DUSE_64_BIT=1")
+            }
+        }
+        stage('Debug') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                CleanBuildAndTest("-DCMAKE_BUILD_TYPE=Debug")
+            }
+        }
+        stage('SIMD fallback') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                CleanBuildAndTest("-DUSE_SIMD_FALLBACK=ON")
+            }
+        }
+        stage('Release [clang]') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                CleanBuildAndTest("-DCMAKE_CXX_FLAGS=-DNVALGRIND")
+                recordIssues(tools: [clang()])
+                sh '''
+                cd build
+                cpack -G DEB
+                cpack -G RPM
+                '''
+                sh '''
+                cd paper
+                make
+                '''
+                minio bucket: 'jenkins', credentialsId: 'jenkins-minio', excludes: '', host: 'minio.weaselab.dev', includes: 'build/*.deb,build/*.rpm,paper/*.pdf', targetFolder: '${JOB_NAME}/${BUILD_NUMBER}/${STAGE_NAME}/'
+            }
+        }
+        stage('gcc') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                CleanBuildAndTest("-DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++")
+                recordIssues(tools: [gcc()])
+            }
+        }
+        stage('Release [clang,aarch64]') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                CleanBuildAndTest("-DCMAKE_TOOLCHAIN_FILE=../aarch64-toolchain.cmake -DCMAKE_CXX_FLAGS=-DNVALGRIND")
+                sh '''
+                cd build
+                cpack -G DEB
+                cpack -G RPM
+                '''
+                minio bucket: 'jenkins', credentialsId: 'jenkins-minio', excludes: '', host: 'minio.weaselab.dev', includes: 'build/*.deb,build/*.rpm', targetFolder: '${JOB_NAME}/${BUILD_NUMBER}/${STAGE_NAME}'
+            }
+        }
+        stage('Coverage') {
+            agent {
+                dockerfile {
+                    args '-v /home/jenkins/ccache:/ccache'
+                    reuseNode true
+                }
+            }
+            steps {
+                script {
+                    gcov_args = "-f ConflictSet.cpp -f LongestCommonPrefix.h -f Metrics.h --gcov-executable 'llvm-cov gcov' --exclude-noncode-lines"
+                }
+                CleanBuildAndTest("-DCMAKE_C_FLAGS=--coverage -DCMAKE_CXX_FLAGS=--coverage -DCMAKE_BUILD_TYPE=Debug -DDISABLE_TSAN=ON")
+                sh """
+                gcovr ${gcov_args} --cobertura > build/coverage.xml
+                """
+                recordCoverage qualityGates: [[criticality: 'NOTE', metric: 'MODULE']], tools: [[parser: 'COBERTURA', pattern: 'build/coverage.xml']]
+                sh """
+                gcovr ${gcov_args}
+                gcovr ${gcov_args} --fail-under-line 100 > /dev/null
+                """
+            }
+        }
+    }
+    post {
+        always {
+            emailext mimeType: 'text/html', body: '${SCRIPT, template="groovy-html.template"}', subject: "${env.JOB_NAME} - Build# ${env.BUILD_NUMBER} - ${currentBuild.currentResult}", to: 'andrew@weaselab.dev'
+        }
+    }
+}

LongestCommonPrefix.h

@@ -96,7 +96,7 @@ inline int firstNeqStride(const uint8_t *ap, const uint8_t *bp) {
 // GCOVR_EXCL_START
 #if defined(HAS_AVX) && !defined(__SANITIZE_THREAD__)
 __attribute__((target("avx512f,avx512bw"))) inline int
-longestCommonPrefixLong(const uint8_t *ap, const uint8_t *bp, int cl) {
+longestCommonPrefix(const uint8_t *ap, const uint8_t *bp, int cl) {
   int i = 0;
   int end = cl & ~63;
   while (i < end) {
@@ -123,7 +123,7 @@ __attribute__((target("default")))
 // GCOVR_EXCL_STOP
 
 inline int
-longestCommonPrefixLong(const uint8_t *ap, const uint8_t *bp, int cl) {
+longestCommonPrefix(const uint8_t *ap, const uint8_t *bp, int cl) {
   if (!(cl >= 0)) {
     __builtin_unreachable(); // GCOVR_EXCL_LINE
   }
@@ -183,71 +183,3 @@ longestCommonPrefixLong(const uint8_t *ap, const uint8_t *bp, int cl) {
 
   return i;
 }
-// Use in-bounds native-word loads for short common lengths. The second load
-// overlaps the first; because the first load matched, any mismatch found in
-// the overlap is impossible and the returned absolute index is still correct.
-// Precondition: cl <= 2 * sizeof(uint64_t)
-__attribute__((always_inline)) static inline int
-longestCommonPrefixShort(const uint8_t *ap, const uint8_t *bp, int cl) {
-  assert(cl <= 2 * int(sizeof(uint64_t)));
-  if (cl >= 8) {
-    uint64_t a0;
-    uint64_t b0;
-    memcpy(&a0, ap, sizeof(a0));
-    memcpy(&b0, bp, sizeof(b0));
-    if (uint64_t mismatched = a0 ^ b0) {
-      return std::countr_zero(mismatched) / 8;
-    }
-    if (cl == 8) {
-      return 8;
-    }
-
-    uint64_t a1;
-    uint64_t b1;
-    memcpy(&a1, ap + cl - 8, sizeof(a1));
-    memcpy(&b1, bp + cl - 8, sizeof(b1));
-    if (uint64_t mismatched = a1 ^ b1) {
-      return cl - 8 + std::countr_zero(mismatched) / 8;
-    }
-    return cl;
-  }
-
-  if (cl >= 4) {
-    uint32_t a0;
-    uint32_t b0;
-    memcpy(&a0, ap, sizeof(a0));
-    memcpy(&b0, bp, sizeof(b0));
-    if (uint32_t mismatched = a0 ^ b0) {
-      return std::countr_zero(mismatched) / 8;
-    }
-    if (cl == 4) {
-      return 4;
-    }
-
-    uint32_t a1;
-    uint32_t b1;
-    memcpy(&a1, ap + cl - 4, sizeof(a1));
-    memcpy(&b1, bp + cl - 4, sizeof(b1));
-    if (uint32_t mismatched = a1 ^ b1) {
-      return cl - 4 + std::countr_zero(mismatched) / 8;
-    }
-    return cl;
-  }
-
-  int i = 0;
-  while (i < cl && ap[i] == bp[i]) {
-    ++i;
-  }
-  return i;
-}
-
-__attribute__((always_inline)) static inline int
-longestCommonPrefix(const uint8_t *ap, const uint8_t *bp, int cl) {
-  if (!(cl >= 0)) {
-    __builtin_unreachable(); // GCOVR_EXCL_LINE
-  }
-  if (cl <= 2 * int(sizeof(uint64_t))) {
-    return longestCommonPrefixShort(ap, bp, cl);
-  }
-  return longestCommonPrefixLong(ap, bp, cl);
-}

README.md

@@ -7,10 +7,10 @@ Hardware for all benchmarks is an AMD Ryzen 9 7900 with (2x32GB) 5600MT/s CL28-3
 ```
 $ clang++ --version
 
-Ubuntu clang version 21.1.8 (6ubuntu1)
+Ubuntu clang version 20.0.0 (++20241120082228+86734c857724-1~exp1~20241120202359.554)
 Target: x86_64-pc-linux-gnu
 Thread model: posix
-InstalledDir: /usr/lib/llvm-21/bin
+InstalledDir: /usr/lib/llvm-20/bin
 ```
 
 # Microbenchmark
@@ -19,30 +19,30 @@ InstalledDir: /usr/lib/llvm-21/bin
 
 |               ns/op |                op/s |    err% |          ins/op |          cyc/op |    IPC |         bra/op |   miss% |     total | benchmark
 |--------------------:|--------------------:|--------:|----------------:|----------------:|-------:|---------------:|--------:|----------:|:----------
-|              165.93 |        6,026,520.04 |    0.1% |        3,107.03 |          610.42 |  5.090 |         558.59 |    0.0% |      1.98 | `point reads`
+|              161.29 |        6,200,056.17 |    0.1% |        3,014.03 |          831.04 |  3.627 |         504.59 |    0.0% |      1.93 | `point reads`
-|              163.08 |        6,132,046.57 |    0.1% |        3,036.76 |          599.68 |  5.064 |         539.35 |    0.0% |      1.95 | `prefix reads`
+|              158.32 |        6,316,160.64 |    0.1% |        2,954.16 |          815.80 |  3.621 |         490.17 |    0.0% |      1.89 | `prefix reads`
-|              250.26 |        3,995,834.21 |    0.1% |        3,722.71 |          920.25 |  4.045 |         692.00 |    0.0% |      2.99 | `range reads`
+|              237.39 |        4,212,409.50 |    0.2% |        3,592.41 |        1,233.96 |  2.911 |         629.31 |    0.0% |      2.84 | `range reads`
-|              359.64 |        2,780,581.63 |    0.1% |        4,523.64 |        1,322.63 |  3.420 |         720.22 |    2.1% |      4.30 | `point writes`
+|              442.11 |        2,261,878.94 |    0.0% |        4,450.57 |        2,314.25 |  1.923 |         707.92 |    2.1% |      5.28 | `point writes`
-|              358.20 |        2,791,726.93 |    0.1% |        4,484.57 |        1,317.23 |  3.405 |         705.00 |    2.0% |      4.28 | `prefix writes`
+|              439.89 |        2,273,308.53 |    0.1% |        4,410.22 |        2,302.29 |  1.916 |         694.74 |    2.1% |      5.25 | `prefix writes`
-|              193.91 |        5,157,081.29 |    0.1% |        2,224.10 |          713.33 |  3.118 |         377.17 |    3.3% |      2.32 | `range writes`
+|              290.96 |        3,436,936.78 |    0.0% |        2,315.38 |        1,528.68 |  1.515 |         396.69 |    3.3% |      3.49 | `range writes`
-|              422.03 |        2,369,513.38 |    1.1% |        6,855.96 |        1,551.55 |  4.419 |       1,227.82 |    1.4% |      0.05 | `monotonic increasing point writes`
+|              476.93 |        2,096,762.02 |    0.6% |        6,999.33 |        2,484.94 |  2.817 |       1,251.73 |    1.3% |      0.06 | `monotonic increasing point writes`
-|          142,467.00 |            7,019.17 |    3.8% |      812,047.33 |      508,535.40 |  1.597 |     151,246.50 |    1.0% |      0.01 | `worst case for radix tree`
+|          131,736.57 |            7,590.91 |    1.1% |      807,444.50 |      704,941.71 |  1.145 |     144,584.60 |    0.9% |      0.01 | `worst case for radix tree`
-|               37.66 |       26,554,217.74 |    0.8% |          701.00 |          138.80 |  5.050 |         102.00 |    0.0% |      0.01 | `create and destroy`
+|               45.50 |       21,978,369.95 |    1.1% |          902.00 |          232.36 |  3.882 |         132.00 |    0.0% |      0.01 | `create and destroy`
 
 ## Radix tree (this implementation)
 
 
 |               ns/op |                op/s |    err% |          ins/op |          cyc/op |    IPC |         bra/op |   miss% |     total | benchmark
 |--------------------:|--------------------:|--------:|----------------:|----------------:|-------:|---------------:|--------:|----------:|:----------
-|               12.25 |       81,664,669.29 |    0.4% |          233.49 |           44.94 |  5.196 |          31.23 |    0.7% |      0.15 | `point reads`
+|               12.36 |       80,885,626.43 |    0.2% |          243.56 |           63.62 |  3.828 |          31.07 |    0.6% |      0.15 | `point reads`
-|               13.35 |       74,903,950.49 |    0.1% |          271.75 |           49.10 |  5.534 |          40.21 |    0.4% |      0.16 | `prefix reads`
+|               14.18 |       70,502,196.81 |    0.1% |          297.72 |           73.13 |  4.071 |          40.31 |    0.5% |      0.17 | `prefix reads`
-|               31.37 |       31,874,583.62 |    0.1% |          700.39 |          115.43 |  6.067 |         100.83 |    0.2% |      0.37 | `range reads`
+|               33.44 |       29,901,623.04 |    0.1% |          767.90 |          172.42 |  4.454 |         101.32 |    0.2% |      0.40 | `range reads`
-|               19.55 |       51,153,057.80 |    0.1% |          366.51 |           71.90 |  5.098 |          47.92 |    0.5% |      0.23 | `point writes`
+|               19.48 |       51,342,564.70 |    0.3% |          374.45 |          100.43 |  3.728 |          48.92 |    0.5% |      0.23 | `point writes`
-|               36.92 |       27,088,344.17 |    0.1% |          645.11 |          135.72 |  4.753 |         100.31 |    0.4% |      0.44 | `prefix writes`
+|               37.46 |       26,694,471.44 |    0.1% |          672.00 |          193.14 |  3.479 |         101.28 |    0.3% |      0.45 | `prefix writes`
-|               36.84 |       27,143,553.08 |    0.1% |          691.04 |          135.42 |  5.103 |         114.01 |    0.1% |      0.45 | `range writes`
+|               38.78 |       25,784,784.34 |    0.0% |          738.26 |          199.93 |  3.693 |         111.59 |    0.1% |      0.47 | `range writes`
-|               74.62 |       13,400,467.72 |    0.9% |        1,418.37 |          274.20 |  5.173 |         273.39 |    0.1% |      0.01 | `monotonic increasing point writes`
+|               76.05 |       13,148,995.74 |    0.7% |        1,450.77 |          397.16 |  3.653 |         275.72 |    0.0% |      0.01 | `monotonic increasing point writes`
-|          352,671.00 |            2,835.50 |   15.3% |    3,687,535.00 |    1,287,119.00 |  2.865 |     712,030.00 |    0.1% |      0.01 | :wavy_dash: `worst case for radix tree` (Unstable with ~1.0 iters. Increase `minEpochIterations` to e.g. 10)
+|          286,920.33 |            3,485.29 |    0.4% |    4,117,948.00 |    1,521,352.00 |  2.707 |     714,833.00 |    0.1% |      0.01 | `worst case for radix tree`
-|               96.63 |       10,348,729.42 |    1.2% |        1,766.00 |          355.53 |  4.967 |         283.00 |    0.0% |      0.01 | `create and destroy`
+|               95.66 |       10,453,798.72 |    0.5% |        1,986.00 |          495.04 |  4.012 |         315.00 |    0.0% |      0.01 | `create and destroy`
 
 # "Real data" test
 
@@ -51,13 +51,13 @@ Point queries only. Gc ratio is the ratio of time spent doing garbage collection
 ## skip list
 
 ```
-Check: 4.69875 seconds, 347.017 MB/s, Add: 3.39475 seconds, 165.153 MB/s, Gc ratio: 38.7414%, Peak idle memory: 5.51852e+06
+Check: 4.53508 seconds, 371.81 MB/s, Add: 3.81222 seconds, 150.919 MB/s, Gc ratio: 33.66%, Peak idle memory: 5.61007e+06
 ```
 
 ## radix tree
 
 ```
-Check: 0.976745 seconds, 1669.37 MB/s, Add: 1.20464 seconds, 465.411 MB/s, Gc ratio: 39.4158%, Peak idle memory: 2.0226e+06
+Check: 0.957735 seconds, 1760.6 MB/s, Add: 1.19942 seconds, 479.678 MB/s, Gc ratio: 38.6069%, Peak idle memory: 2.05667e+06
 ```
 
 ## hash table
@@ -65,6 +65,6 @@ Check: 0.976745 seconds, 1669.37 MB/s, Add: 1.20464 seconds, 465.411 MB/s, Gc ra
 (The hash table implementation doesn't work on range queries, and its purpose is to provide an idea of how fast point queries can be)
 
 ```
-Check: 0.864589 seconds, 1885.92 MB/s, Add: 0.634604 seconds, 883.47 MB/s, Gc ratio: 41.184%, Peak idle memory: 0
+Check: 0.804598 seconds, 2095.69 MB/s, Add: 0.671221 seconds, 857.147 MB/s, Gc ratio: 35.0034%, Peak idle memory: 0
 ```
 

RealDataBench.cpp

@@ -1,6 +1,5 @@
 #include <ConflictSet.h>
 
-#include <algorithm>
 #include <cerrno>
 #include <chrono>
 #include <cstdio>
@@ -78,10 +77,10 @@ int main(int argc, const char **argv) {
       begin = end + 1;
       end = (uint8_t *)memchr(begin, '\n', size);
 
-      if (line.size() >= 2 && line[0] == 'P') {
+      if (line.size() > 0 && line[0] == 'P') {
-        write = line.subspan(2, line.size() - 2);
+        write = line.subspan(2, line.size());
-      } else if (line.size() >= 2 && line[0] == 'L') {
+      } else if (line.size() > 0 && line[0] == 'L') {
-        reads.push_back(line.subspan(2, line.size() - 2));
+        reads.push_back(line.subspan(2, line.size()));
       } else if (line.empty()) {
         {
           readRanges.resize(reads.size());
@@ -91,7 +90,7 @@ int main(int argc, const char **argv) {
             iter->begin.len = read.size();
             checkBytes += read.size();
             iter->end.len = 0;
-            iter->readVersion = std::max<int64_t>(0, version - 100);
+            iter->readVersion = version - 100;
             ++iter;
           }
         }
@@ -122,7 +121,7 @@ int main(int argc, const char **argv) {
         }
 
         timer = now();
-        cs.setOldestVersion(std::max<int64_t>(0, version - 10000));
+        cs.setOldestVersion(version - 10000);
         gcTime += now() - timer;
       }
     }

ServerBench.cpp

@@ -12,7 +12,6 @@
 #include <sys/ioctl.h>
 #include <sys/resource.h>
 #include <sys/socket.h>
-#include <sys/syscall.h>
 #include <sys/types.h>
 #include <sys/uio.h>
 #include <thread>

ShowMemoryTest.cpp

@@ -1,115 +0,0 @@
-// Exercises the SHOW_MEMORY instrumentation, which the other tests don't
-// compile with (this target passes -DSHOW_MEMORY=1). The instrumentation has
-// its own consistency checks, which abort on failure, and asserts are enabled
-// by TEST_FLAGS. Only one ConflictSet may exist at a time, since the
-// accounting is process-wide.
-#include "ConflictSet.h"
-#include "Internal.h"
-
-#include <cassert>
-#include <cstdint>
-#include <string>
-
-using namespace weaselab;
-
-namespace {
-
-void writePoint(ConflictSet &cs, const std::string &key, int64_t writeVersion) {
-  ConflictSet::WriteRange w{{(const uint8_t *)key.data(), int(key.size())},
-                            {nullptr, 0}};
-  cs.addWrites(&w, 1, writeVersion);
-}
-
-ConflictSet::Result checkPoint(const ConflictSet &cs, const std::string &key,
-                               int64_t readVersion) {
-  ConflictSet::ReadRange r{{(const uint8_t *)key.data(), int(key.size())},
-                           {nullptr, 0},
-                           readVersion};
-  ConflictSet::Result result;
-  cs.check(&r, &result, 1);
-  return result;
-}
-
-// Everything a ConflictSet allocated must have been accounted for and released
-// by the time it's destroyed.
-void checkFullyReleased() {
-  assert(nodeBytes == 0);
-  assert(partialCapacityBytes == 0);
-  assert(totalKeys == 0);
-  assert(keyBytes == 0);
-  assert(mallocBytes == 0);
-}
-
-} // namespace
-
-int main() {
-  // Splitting a node replaces it, which defers the release of the old node.
-  // Releasing a deferred node must account for its capacity, and the key it
-  // held must be accounted for without walking `parent` pointers of nodes that
-  // have already been released.
-  {
-    ConflictSet cs{0};
-    writePoint(cs, "a", 1);
-    writePoint(cs, "ab", 2);
-    assert(checkPoint(cs, "a", 0) == ConflictSet::Conflict);
-    assert(checkPoint(cs, "ab", 0) == ConflictSet::Conflict);
-    assert(checkPoint(cs, "b", 0) == ConflictSet::Commit);
-    assert(nodeBytes > 0);
-    assert(keyBytes > 0);
-  }
-  checkFullyReleased();
-
-  // Exceeding the version window erases the tree mid-workload.
-  {
-    ConflictSet cs{0};
-    writePoint(cs, "abc", 1);
-    const int64_t version = int64_t(1) << 40;
-    writePoint(cs, "abc", version);
-    assert(checkPoint(cs, "abc", 0) == ConflictSet::TooOld);
-    assert(checkPoint(cs, "abc", version) == ConflictSet::Commit);
-  }
-  checkFullyReleased();
-
-  // Erase a tree with many nodes, both mid-workload and at destruction.
-  {
-    ConflictSet cs{0};
-    for (int i = 0; i < 1000; ++i) {
-      writePoint(cs, "prefix/" + std::to_string(i), i + 1);
-    }
-    const int64_t version = int64_t(1) << 40;
-    writePoint(cs, "prefix/1", version);
-    assert(checkPoint(cs, "prefix/1", version) == ConflictSet::Commit);
-
-    for (int i = 0; i < 1000; ++i) {
-      writePoint(cs, "other/" + std::to_string(i), version + i + 1);
-    }
-    assert(checkPoint(cs, "other/999", version) == ConflictSet::Conflict);
-  }
-  checkFullyReleased();
-
-  // Grow a node through every node type, so that the accounting for each node
-  // type's size is exercised, then erase the entries one at a time so that
-  // nodes are released individually rather than as part of a whole tree.
-  {
-    ConflictSet cs{0};
-    std::string key(1, '\0');
-    for (int i = 0; i < 256; ++i) {
-      key[0] = char(i);
-      writePoint(cs, key, i + 1);
-    }
-    for (int i = 0; i < 256; ++i) {
-      key[0] = char(i);
-      assert(checkPoint(cs, key, i) == ConflictSet::Conflict);
-    }
-
-    // Garbage collection erases entries that can only fail with TooOld anyway.
-    // Doing it one version at a time walks the tree, erasing, downsizing and
-    // releasing nodes as it goes.
-    for (int i = 0; i < 256; ++i) {
-      cs.setOldestVersion(i + 1);
-    }
-    key[0] = '\0';
-    assert(checkPoint(cs, key, 0) == ConflictSet::TooOld);
-  }
-  checkFullyReleased();
-}

SkipList.cpp

@@ -38,9 +38,6 @@ std::span<const uint8_t> keyAfter(Arena &arena, std::span<const uint8_t> key) {
 
 std::span<const uint8_t> copyToArena(Arena &arena,
                                      std::span<const uint8_t> key) {
-  if (key.size() == 0) {
-    return std::span<const uint8_t>();
-  }
   auto result = std::span<uint8_t>(new (arena) uint8_t[key.size()], key.size());
   memcpy(result.data(), key.data(), key.size());
   return result;
@@ -100,7 +97,7 @@ force_inline bool getCharacter(const KeyInfo &ki, int character,
 
 bool operator<(const KeyInfo &lhs, const KeyInfo &rhs) {
   int i = std::min(lhs.key.size(), rhs.key.size());
-  int c = i > 0 ? memcmp(lhs.key.data(), rhs.key.data(), i) : 0;
+  int c = memcmp(lhs.key.data(), rhs.key.data(), i);
   if (c != 0)
     return c < 0;
 
@@ -290,8 +287,7 @@ private:
 
   static force_inline bool less(const uint8_t *a, int aLen, const uint8_t *b,
                                 int bLen) {
-    int n = std::min(aLen, bLen);
+    int c = memcmp(a, b, std::min(aLen, bLen));
-    int c = n > 0 ? memcmp(a, b, n) : 0;
     if (c < 0)
       return true;
     if (c > 0)
@@ -758,8 +754,7 @@ struct __attribute__((visibility("hidden"))) ConflictSet::Impl {
 
   void addWrites(const ConflictSet::WriteRange *writes, int count,
                  int64_t writeVersion) {
-    auto points = std::vector<KeyInfo>();
+    auto points = std::vector<KeyInfo>(count * 2);
-    points.reserve(count * 2);
     Arena arena;
 
     for (int r = 0; r < count; r++) {
@@ -819,17 +814,10 @@ struct __attribute__((visibility("hidden"))) ConflictSet::Impl {
   }
 
   void setOldestVersion(int64_t oldestVersion) {
-    // oldestVersion becomes the max of its previous value and the argument, so
-    // an older version is a no-op. We still fall through in the equal case so
-    // that garbage collection can make progress without advancing the oldest
-    // version.
-    if (oldestVersion < this->oldestVersion) {
-      return;
-    }
-
     // This isn't 100% accurate. It overcounts if you hit the end
     gc_iterations_total.add(keyUpdates);
 
+    assert(oldestVersion >= this->oldestVersion);
     this->oldestVersion = oldestVersion;
     SkipList::Finger finger;
     int temp;
@@ -993,12 +981,7 @@ ConflictSet::ConflictSet(ConflictSet &&other) noexcept
     : impl(std::exchange(other.impl, nullptr)) {}
 
 ConflictSet &ConflictSet::operator=(ConflictSet &&other) noexcept {
-  if (this != &other) {
-    if (impl) {
-      internal_destroy(impl);
-    }
   impl = std::exchange(other.impl, nullptr);
-  }
   return *this;
 }
 

TestDriver.cpp

@@ -35,5 +35,36 @@ int main(int argc, char **argv) {
         break;
       }
     }
+
+    {
+      ConflictSet::MetricsV1 *metrics;
+      int metricsCount;
+      driver1.cs.getMetricsV1(&metrics, &metricsCount);
+      printf("#################### METRICS for ConflictSet 1 for %s "
+             "####################\n",
+             argv[i]);
+      for (int i = 0; i < metricsCount; ++i) {
+        printf("# HELP %s %s\n", metrics[i].name, metrics[i].help);
+        printf("# TYPE %s %s\n", metrics[i].name,
+               metrics[i].type == metrics[i].Counter ? "counter" : "gauge");
+        printf("%s %g\n", metrics[i].name, metrics[i].getValue());
+      }
+      puts("");
+    }
+    {
+      ConflictSet::MetricsV1 *metrics;
+      int metricsCount;
+      driver2.cs.getMetricsV1(&metrics, &metricsCount);
+      printf("#################### METRICS for ConflictSet 2 for %s "
+             "####################\n",
+             argv[i]);
+      for (int i = 0; i < metricsCount; ++i) {
+        printf("# HELP %s %s\n", metrics[i].name, metrics[i].help);
+        printf("# TYPE %s %s\n", metrics[i].name,
+               metrics[i].type == metrics[i].Counter ? "counter" : "gauge");
+        printf("%s %g\n", metrics[i].name, metrics[i].getValue());
+      }
+      puts("");
+    }
   }
 }

build_msan_toolchain.sh

@@ -1,43 +0,0 @@
-#!/bin/bash
-
-set -euxo pipefail
-
-LLVM_VERSION="${LLVM_VERSION:-21}"
-MSAN_PREFIX="${MSAN_PREFIX:-$PWD/msan}"
-JOBS="${JOBS:-$(nproc)}"
-
-cd /tmp
-rm -rf libcxx-msan
-mkdir libcxx-msan
-cd libcxx-msan
-
-git clone --depth=1 "https://github.com/llvm/llvm-project.git" -b "release/${LLVM_VERSION}.x"
-
-cmake -S llvm-project/runtimes -B build_msan \
-  -DCMAKE_BUILD_TYPE=Release \
-  -DCMAKE_CXX_COMPILER="clang++-${LLVM_VERSION}" \
-  -DCMAKE_C_COMPILER="clang-${LLVM_VERSION}" \
-  -DCMAKE_INSTALL_PREFIX="${MSAN_PREFIX}" \
-  -DLLVM_ENABLE_RUNTIMES="libcxx;libcxxabi;libunwind" \
-  -DLLVM_USE_SANITIZER=MemoryWithOrigins
-
-cmake --build build_msan -j"${JOBS}"
-cmake --install build_msan
-
-cmake -S llvm-project/runtimes -B build_nomsan \
-  -DCMAKE_BUILD_TYPE=Release \
-  -DCMAKE_CXX_COMPILER="clang++-${LLVM_VERSION}" \
-  -DCMAKE_C_COMPILER="clang-${LLVM_VERSION}" \
-  -DCMAKE_INSTALL_PREFIX="${MSAN_PREFIX}" \
-  -DLLVM_ENABLE_RUNTIMES="libunwind"
-
-cmake --build build_nomsan -j"${JOBS}"
-cmake --install build_nomsan
-
-VERSION="$(clang-"${LLVM_VERSION}" --version | head -n1 | sed -E 's/.*clang version ([0-9.]+).*/\1/')"
-TARBALL="msan-toolchain-${VERSION}.tar.zst"
-
-tar --zstd -cf "${TARBALL}" -C "${MSAN_PREFIX}" .
-
-TARBALL_PATH="$(pwd)/${TARBALL}"
-echo "Created: ${TARBALL_PATH}"

conflict_set.py

@@ -27,37 +27,23 @@ class Result(enum.Enum):
     TOO_OLD = 2
 
 
-def _make_key(buf: bytes) -> tuple[_Key, bytearray]:
-    """Create a _Key and a backing bytearray that must be kept alive."""
-    backing = bytearray(buf)
-    array = (ctypes.c_ubyte * len(backing)).from_buffer(backing)
-    return _Key(array, len(array)), backing
-
-
 def write(begin: bytes, end: Optional[bytes] = None) -> WriteRange:
-    begin_key, begin_buf = _make_key(begin)
+    b = (ctypes.c_ubyte * len(begin)).from_buffer(bytearray(begin))
+
     if end is None:
-        end_key = _Key((ctypes.c_ubyte * 0)(), 0)
+        e = (ctypes.c_ubyte * 0)()
-        end_buf = None
     else:
-        end_key, end_buf = _make_key(end)
+        e = (ctypes.c_ubyte * len(end)).from_buffer(bytearray(end))
-    result = WriteRange(begin_key, end_key)
+    return WriteRange(_Key(b, len(b)), _Key(e, len(e)))
-    result._begin_buf = begin_buf
-    result._end_buf = end_buf
-    return result
 
 
 def read(version: int, begin: bytes, end: Optional[bytes] = None) -> ReadRange:
-    begin_key, begin_buf = _make_key(begin)
+    b = (ctypes.c_ubyte * len(begin)).from_buffer(bytearray(begin))
     if end is None:
-        end_key = _Key((ctypes.c_ubyte * 0)(), 0)
+        e = (ctypes.c_ubyte * 0)()
-        end_buf = None
     else:
-        end_key, end_buf = _make_key(end)
+        e = (ctypes.c_ubyte * len(end)).from_buffer(bytearray(end))
-    result = ReadRange(begin_key, end_key, version)
+    return ReadRange(_Key(b, len(b)), _Key(e, len(e)), version)
-    result._begin_buf = begin_buf
-    result._end_buf = end_buf
-    return result
 
 
 class ConflictSet:
@@ -102,7 +88,6 @@ class ConflictSet:
             ctypes.POINTER(ctypes.c_int),
             ctypes.c_int,
         )
-        self._lib.ConflictSet_check.restype = None
 
         self._lib.ConflictSet_addWrites.argtypes = (
             ctypes.c_void_p,
@@ -110,16 +95,13 @@ class ConflictSet:
             ctypes.c_int,
             ctypes.c_int64,
         )
-        self._lib.ConflictSet_addWrites.restype = None
 
         self._lib.ConflictSet_setOldestVersion.argtypes = (
             ctypes.c_void_p,
             ctypes.c_int64,
         )
-        self._lib.ConflictSet_setOldestVersion.restype = None
 
         self._lib.ConflictSet_destroy.argtypes = (ctypes.c_void_p,)
-        self._lib.ConflictSet_destroy.restype = None
 
         self._lib.ConflictSet_getBytes.argtypes = (ctypes.c_void_p,)
         self._lib.ConflictSet_getBytes.restype = ctypes.c_int64

conflict_set_cxx_api_test.cpp

@@ -2,7 +2,6 @@
 
 #include <cassert>
 #include <cstdio>
-#include <limits>
 
 using namespace weaselab;
 
@@ -24,28 +23,6 @@ int main(void) {
   int64_t bytes = cs.getBytes();
   assert(bytes > 0);
 
-  // Regression: a write just below INT64_MAX must not overflow the nominal
-  // version window comparison and erase entries that reads still observe.
-  // Versions are only required to be >= 0, so this is valid input.
-  {
-    const int64_t imax = std::numeric_limits<int64_t>::max();
-    const int64_t v0 = imax - int64_t(1000000000);
-    ConflictSet nearMax(0);
-    ConflictSet::WriteRange w2;
-    w2.begin.p = (const uint8_t *)"key";
-    w2.begin.len = 3;
-    w2.end.len = 0;
-    nearMax.addWrites(&w2, 1, v0);
-    nearMax.addWrites(&w2, 0, imax);
-    ConflictSet::ReadRange r2;
-    r2.begin.p = (const uint8_t *)"key";
-    r2.begin.len = 3;
-    r2.end.len = 0;
-    r2.readVersion = v0 - 1;
-    nearMax.check(&r2, &result, 1);
-    assert(result == ConflictSet::Conflict);
-  }
-
   ConflictSet::MetricsV1 *metrics;
   int metricsCount;
   cs.getMetricsV1(&metrics, &metricsCount);