Bound memory, and disable free list for now

CC #9

.clangd

@@ -1,2 +1,2 @@
 CompileFlags:
-  Add: [-DENABLE_MAIN, -UNDEBUG, -DENABLE_FUZZ, -DTHREAD_TEST, -fexceptions]
+  Add: [-DENABLE_MAIN, -UNDEBUG, -DENABLE_FUZZ, -fexceptions]

.gitignore

@@ -1,3 +1,2 @@
 .cache
-__pycache__
 build

.pre-commit-config.yaml

@@ -1,11 +1,11 @@
 repos:
   - repo: https://github.com/pre-commit/mirrors-clang-format
-    rev: 6d365699efc33b1b432eab5b4ae331a19e1857de # frozen: v18.1.2
+    rev: b111689e7b5cba60be3c62d5db2bd1357f4d36ca
     hooks:
       - id: clang-format
         exclude: ".*third_party/.*"
   - repo: https://github.com/cheshirekow/cmake-format-precommit
-    rev: e2c2116d86a80e72e7146a06e68b7c228afc6319 # frozen: v0.6.13
+    rev: e2c2116d86a80e72e7146a06e68b7c228afc6319
     hooks:
       - id: cmake-format
   - repo: local
@@ -13,7 +13,7 @@ repos:
       - id: debug verbose check
         name: disallow checking in DEBUG_VERBOSE=1
         description: disallow checking in DEBUG_VERBOSE=1
-        entry: "^#define DEBUG_VERBOSE 1$"
+        entry: '^#define DEBUG_VERBOSE 1$'
         language: pygrep
         types: [c++]
   - repo: local
@@ -21,14 +21,6 @@ repos:
       - id: debug verbose check
         name: disallow checking in SHOW_MEMORY=1
         description: disallow checking in SHOW_MEMORY=1
-        entry: "^#define SHOW_MEMORY 1$"
+        entry: '^#define SHOW_MEMORY 1$'
         language: pygrep
         types: [c++]
-  - repo: https://github.com/shellcheck-py/shellcheck-py
-    rev: a23f6b85d0fdd5bb9d564e2579e678033debbdff # frozen: v0.10.0.1
-    hooks:
-      - id: shellcheck
-  - repo: https://github.com/psf/black
-    rev: 552baf822992936134cbd31a38f69c8cfe7c0f05 # frozen: 24.3.0
-    hooks:
-      - id: black

Bench.cpp

@@ -34,7 +34,7 @@ ConflictSet::ReadRange singleton(Arena &arena, std::span<const uint8_t> key) {
       std::span<uint8_t>(new (arena) uint8_t[key.size() + 1], key.size() + 1);
   memcpy(r.data(), key.data(), key.size());
   r[key.size()] = 0;
-  return {{key.data(), int(key.size())}, {r.data(), int(r.size())}, 0};
+  return {key.data(), int(key.size()), r.data(), int(r.size())};
 }
 
 ConflictSet::ReadRange prefixRange(Arena &arena, std::span<const uint8_t> key) {
@@ -52,7 +52,7 @@ ConflictSet::ReadRange prefixRange(Arena &arena, std::span<const uint8_t> key) {
   auto r = std::span<uint8_t>(new (arena) uint8_t[index + 1], index + 1);
   memcpy(r.data(), key.data(), index + 1);
   r[r.size() - 1]++;
-  return {{key.data(), int(key.size())}, {r.data(), int(r.size())}, 0};
+  return {key.data(), int(key.size()), r.data(), int(r.size())};
 }
 
 void benchConflictSet() {

CMakeLists.txt

@@ -1,28 +1,15 @@
 cmake_minimum_required(VERSION 3.18)
 project(
-  conflict-set
+  conflict_set
-  VERSION 0.0.4
+  VERSION 0.0.1
   DESCRIPTION
     "A data structure for optimistic concurrency control on ranges of bitwise-lexicographically-ordered keys."
   HOMEPAGE_URL "https://git.weaselab.dev/weaselab/conflict-set"
   LANGUAGES C CXX)
 set(CMAKE_CXX_STANDARD 20)
 
-file(WRITE ${CMAKE_BINARY_DIR}/version.txt ${PROJECT_VERSION})
-
-include(CMakePushCheckState)
-include(CheckCXXCompilerFlag)
-include(CheckIncludeFileCXX)
-include(CheckCXXSourceCompiles)
-
 set(DEFAULT_BUILD_TYPE "Release")
 
-if(EMSCRIPTEN OR CMAKE_SYSTEM_NAME STREQUAL WASI)
-  set(WASM ON)
-else()
-  set(WASM OFF)
-endif()
-
 if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
   message(
     STATUS
@@ -35,33 +22,16 @@ if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
                                                "MinSizeRel" "RelWithDebInfo")
 endif()
 
-add_compile_options(-fdata-sections -ffunction-sections -Wswitch-enum
+add_compile_options(-fdata-sections -ffunction-sections)
-                    -Werror=switch-enum -fPIC)
-
-set(full_relro_flags "-pie;LINKER:-z,relro,-z,now,-z,noexecstack")
-cmake_push_check_state()
-list(APPEND CMAKE_REQUIRED_LINK_OPTIONS ${full_relro_flags})
-check_cxx_source_compiles("int main(){}" HAS_FULL_RELRO FAIL_REGEX "warning:")
-if(HAS_FULL_RELRO)
-  add_link_options(${full_relro_flags})
-endif()
-cmake_pop_check_state()
-
-set(version_script_flags LINKER:--version-script=${CMAKE_SOURCE_DIR}/linker.map)
-cmake_push_check_state()
-list(APPEND CMAKE_REQUIRED_LINK_OPTIONS ${version_script_flags})
-check_cxx_source_compiles("int main(){}" HAS_VERSION_SCRIPT FAIL_REGEX
-                          "warning:")
-cmake_pop_check_state()
-
-option(USE_SIMD_FALLBACK
-       "Use fallback implementations of functions that use SIMD" OFF)
 
 # This is encouraged according to
 # https://valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.clientreq
 include_directories(SYSTEM ${CMAKE_SOURCE_DIR}/third_party/valgrind)
 
-add_compile_options($<$<COMPILE_LANGUAGE:CXX>:-Wno-invalid-offsetof>)
+if(CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+  add_compile_options(-Wno-maybe-uninitialized
+                      $<$<COMPILE_LANGUAGE:CXX>:-Wno-invalid-offsetof>)
+endif()
 
 if(APPLE)
   add_link_options(-Wl,-dead_strip)
@@ -69,46 +39,32 @@ else()
   add_link_options(-Wl,--gc-sections)
 endif()
 
-if(EMSCRIPTEN)
+include(CheckIncludeFileCXX)
-  # https://github.com/emscripten-core/emscripten/issues/15377#issuecomment-1285167486
+include(CMakePushCheckState)
-  add_link_options(-lnodefs.js -lnoderawfs.js)
+
-  add_link_options(-s ALLOW_MEMORY_GROWTH)
+cmake_push_check_state()
+list(APPEND CMAKE_REQUIRED_FLAGS -mavx)
+check_include_file_cxx("immintrin.h" HAS_AVX)
+if(HAS_AVX)
+  add_compile_options(-mavx)
+  add_compile_definitions(HAS_AVX)
 endif()
+cmake_pop_check_state()
 
-if(NOT USE_SIMD_FALLBACK)
+check_include_file_cxx("arm_neon.h" HAS_ARM_NEON)
-  cmake_push_check_state()
+if(HAS_ARM_NEON)
-  list(APPEND CMAKE_REQUIRED_FLAGS -msimd128)
+  add_compile_definitions(HAS_ARM_NEON)
-  check_include_file_cxx("wasm_simd128.h" HAS_WASM_SIMD)
-  if(HAS_WASM_SIMD)
-    add_compile_options(-msimd128)
-    add_compile_definitions(HAS_WASM_SIMD)
-  endif()
-  cmake_pop_check_state()
-
-  cmake_push_check_state()
-  list(APPEND CMAKE_REQUIRED_FLAGS -mavx)
-  check_include_file_cxx("immintrin.h" HAS_AVX)
-  if(HAS_AVX)
-    add_compile_options(-mavx)
-    add_compile_definitions(HAS_AVX)
-  endif()
-  cmake_pop_check_state()
-
-  check_include_file_cxx("arm_neon.h" HAS_ARM_NEON)
-  if(HAS_ARM_NEON)
-    add_compile_definitions(HAS_ARM_NEON)
-  endif()
 endif()
 
 set(CMAKE_CXX_IMPLICIT_LINK_LIBRARIES "")
 
-add_library(${PROJECT_NAME}-object OBJECT ConflictSet.cpp)
+add_library(${PROJECT_NAME}_object OBJECT ConflictSet.cpp)
-target_compile_options(${PROJECT_NAME}-object PRIVATE -fno-exceptions
+target_compile_options(${PROJECT_NAME}_object PRIVATE -fPIC -fno-exceptions
                                                       -fvisibility=hidden)
-target_include_directories(${PROJECT_NAME}-object
+target_include_directories(${PROJECT_NAME}_object
                            PRIVATE ${CMAKE_SOURCE_DIR}/include)
 
-add_library(${PROJECT_NAME} SHARED $<TARGET_OBJECTS:${PROJECT_NAME}-object>)
+add_library(${PROJECT_NAME} SHARED $<TARGET_OBJECTS:${PROJECT_NAME}_object>)
 set_target_properties(
   ${PROJECT_NAME} PROPERTIES LIBRARY_OUTPUT_DIRECTORY
                              "${CMAKE_BINARY_DIR}/radix_tree")
@@ -116,32 +72,24 @@ if(NOT CMAKE_BUILD_TYPE STREQUAL Debug)
   set_target_properties(${PROJECT_NAME} PROPERTIES LINKER_LANGUAGE C)
 endif()
 
-if(HAS_VERSION_SCRIPT)
+if(NOT APPLE)
   target_link_options(${PROJECT_NAME} PRIVATE
                       LINKER:--version-script=${CMAKE_SOURCE_DIR}/linker.map)
 endif()
 
-add_library(${PROJECT_NAME}-static STATIC
+add_library(${PROJECT_NAME}_static STATIC
-            $<TARGET_OBJECTS:${PROJECT_NAME}-object>)
+            $<TARGET_OBJECTS:${PROJECT_NAME}_object>)
 if(NOT CMAKE_BUILD_TYPE STREQUAL Debug)
-  set_target_properties(${PROJECT_NAME}-static PROPERTIES LINKER_LANGUAGE C)
+  set_target_properties(${PROJECT_NAME}_static PROPERTIES LINKER_LANGUAGE C)
 endif()
 
-if(APPLE)
+if(NOT APPLE AND CMAKE_OBJCOPY)
   add_custom_command(
-    TARGET ${PROJECT_NAME}-static
+    TARGET conflict_set_static
-    PRE_LINK
-    COMMAND ${CMAKE_SOURCE_DIR}/privatize_symbols_macos.sh
-            $<TARGET_OBJECTS:${PROJECT_NAME}-object>)
-else()
-  add_custom_command(
-    TARGET ${PROJECT_NAME}-static
     POST_BUILD
     COMMAND
-      ${CMAKE_OBJCOPY}
+      ${CMAKE_OBJCOPY} --keep-global-symbols=${CMAKE_SOURCE_DIR}/symbols.txt
-      --keep-global-symbols=${CMAKE_SOURCE_DIR}/symbol-exports.txt
+      $<TARGET_FILE:${PROJECT_NAME}_static>)
-      $<TARGET_FILE:${PROJECT_NAME}-static> || echo
-      "Proceeding with all symbols global in static library")
 endif()
 
 set(TEST_FLAGS -Wall -Wextra -Wunreachable-code -Wpedantic -UNDEBUG)
@@ -150,51 +98,30 @@ include(CTest)
 
 if(BUILD_TESTING)
 
-  # corpus tests, which are tests curated by libfuzzer. The goal is to get broad
+  # Shared library version of FoundationDB's skip list implementation
-  # coverage with a small number of tests.
+  add_library(skip_list SHARED SkipList.cpp)
+  target_compile_options(skip_list PRIVATE -fPIC -fno-exceptions
+                                           -fvisibility=hidden)
+  target_include_directories(skip_list PUBLIC ${CMAKE_SOURCE_DIR}/include)
+  set_target_properties(skip_list PROPERTIES LIBRARY_OUTPUT_DIRECTORY
+                                             "${CMAKE_BINARY_DIR}/skip_list")
+  set_target_properties(skip_list PROPERTIES OUTPUT_NAME ${PROJECT_NAME})
+  set_target_properties(skip_list PROPERTIES VERSION ${PROJECT_VERSION}
+                                             SOVERSION ${PROJECT_VERSION_MAJOR})
 
-  file(GLOB CORPUS_TESTS ${CMAKE_SOURCE_DIR}/corpus/*)
+  # Shared library version of a std::unordered_map-based conflict set (point
+  # queries only)
+  add_library(hash_table SHARED HashTable.cpp)
+  target_compile_options(hash_table PRIVATE -fPIC -fno-exceptions
+                                            -fvisibility=hidden)
+  target_include_directories(hash_table PUBLIC ${CMAKE_SOURCE_DIR}/include)
+  set_target_properties(hash_table PROPERTIES LIBRARY_OUTPUT_DIRECTORY
+                                              "${CMAKE_BINARY_DIR}/hash_table")
+  set_target_properties(hash_table PROPERTIES OUTPUT_NAME ${PROJECT_NAME})
+  set_target_properties(
+    hash_table PROPERTIES VERSION ${PROJECT_VERSION} SOVERSION
+                                                     ${PROJECT_VERSION_MAJOR})
 
-  # extra testing that relies on shared libraries, which aren't available with
-  # wasm
-  if(NOT WASM)
-    # Shared library version of FoundationDB's skip list implementation
-    add_library(skip_list SHARED SkipList.cpp)
-    target_compile_options(skip_list PRIVATE -fno-exceptions
-                                             -fvisibility=hidden)
-    target_include_directories(skip_list PUBLIC ${CMAKE_SOURCE_DIR}/include)
-    set_target_properties(skip_list PROPERTIES LIBRARY_OUTPUT_DIRECTORY
-                                               "${CMAKE_BINARY_DIR}/skip_list")
-    set_target_properties(skip_list PROPERTIES OUTPUT_NAME ${PROJECT_NAME})
-    set_target_properties(
-      skip_list PROPERTIES VERSION ${PROJECT_VERSION} SOVERSION
-                                                      ${PROJECT_VERSION_MAJOR})
-
-    # Shared library version of a std::unordered_map-based conflict set (point
-    # queries only)
-    add_library(hash_table SHARED HashTable.cpp)
-    target_compile_options(hash_table PRIVATE -fno-exceptions
-                                              -fvisibility=hidden)
-    target_include_directories(hash_table PUBLIC ${CMAKE_SOURCE_DIR}/include)
-    set_target_properties(
-      hash_table PROPERTIES LIBRARY_OUTPUT_DIRECTORY
-                            "${CMAKE_BINARY_DIR}/hash_table")
-    set_target_properties(hash_table PROPERTIES OUTPUT_NAME ${PROJECT_NAME})
-    set_target_properties(
-      hash_table PROPERTIES VERSION ${PROJECT_VERSION} SOVERSION
-                                                       ${PROJECT_VERSION_MAJOR})
-
-    add_executable(driver_skip_list TestDriver.cpp)
-    target_compile_options(driver_skip_list PRIVATE ${TEST_FLAGS})
-    target_link_libraries(driver_skip_list PRIVATE skip_list)
-
-    foreach(TEST ${CORPUS_TESTS})
-      get_filename_component(hash ${TEST} NAME)
-      add_test(NAME skip_list_${hash} COMMAND driver_skip_list ${TEST})
-    endforeach()
-  endif()
-
-  # ad hoc testing
   add_executable(conflict_set_main ConflictSet.cpp)
   target_include_directories(conflict_set_main
                              PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
@@ -221,13 +148,12 @@ if(BUILD_TESTING)
     endif()
   endif()
 
-  # whitebox tests
+  # corpus tests
+
+  file(GLOB CORPUS_TESTS ${CMAKE_SOURCE_DIR}/corpus/*)
+
   add_executable(fuzz_driver ConflictSet.cpp FuzzTestDriver.cpp)
   target_compile_options(fuzz_driver PRIVATE ${TEST_FLAGS})
-  if(NOT CMAKE_CROSSCOMPILING)
-    target_compile_options(fuzz_driver PRIVATE -fsanitize=address,undefined)
-    target_link_options(fuzz_driver PRIVATE -fsanitize=address,undefined)
-  endif()
   target_compile_definitions(fuzz_driver PRIVATE ENABLE_FUZZ)
   target_include_directories(fuzz_driver
                              PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
@@ -236,47 +162,13 @@ if(BUILD_TESTING)
     add_test(NAME conflict_set_fuzz_${hash} COMMAND fuzz_driver ${TEST})
   endforeach()
 
-  # tsan tests
-  if(NOT CMAKE_CROSSCOMPILING AND NOT CMAKE_BUILD_TYPE STREQUAL Debug)
-    add_executable(tsan_driver ConflictSet.cpp FuzzTestDriver.cpp)
-    target_compile_options(tsan_driver PRIVATE ${TEST_FLAGS} -fsanitize=thread)
-    target_link_options(tsan_driver PRIVATE -fsanitize=thread)
-    target_compile_definitions(tsan_driver PRIVATE ENABLE_FUZZ THREAD_TEST)
-    target_include_directories(tsan_driver
-                               PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include)
-    foreach(TEST ${CORPUS_TESTS})
-      get_filename_component(hash ${TEST} NAME)
-      add_test(NAME conflict_set_tsan_${hash} COMMAND tsan_driver ${TEST})
-    endforeach()
-  endif()
-
-  # blackbox tests
   add_executable(driver TestDriver.cpp)
   target_compile_options(driver PRIVATE ${TEST_FLAGS})
   target_link_libraries(driver PRIVATE ${PROJECT_NAME})
-  foreach(TEST ${CORPUS_TESTS})
-    get_filename_component(hash ${TEST} NAME)
-    add_test(NAME conflict_set_blackbox_${hash} COMMAND driver ${TEST})
-  endforeach()
 
-  # scripted tests. Written manually to fill in anything libfuzzer couldn't
+  add_executable(driver_skip_list TestDriver.cpp)
-  # find.
+  target_compile_options(driver_skip_list PRIVATE ${TEST_FLAGS})
-  if(NOT CMAKE_CROSSCOMPILING)
+  target_link_libraries(driver_skip_list PRIVATE skip_list)
-    find_package(Python3 REQUIRED COMPONENTS Interpreter)
-    set_property(
-      DIRECTORY
-      APPEND
-      PROPERTY CMAKE_CONFIGURE_DEPENDS ${CMAKE_SOURCE_DIR}/test_conflict_set.py)
-    execute_process(
-      COMMAND ${Python3_EXECUTABLE} ${CMAKE_SOURCE_DIR}/test_conflict_set.py
-              list OUTPUT_VARIABLE SCRIPT_TESTS)
-    foreach(TEST ${SCRIPT_TESTS})
-      add_test(
-        NAME script_test_${TEST}
-        COMMAND ${Python3_EXECUTABLE} ${CMAKE_SOURCE_DIR}/test_conflict_set.py
-                test ${TEST} --build-dir ${CMAKE_BINARY_DIR})
-    endforeach()
-  endif()
 
   find_program(VALGRIND_EXE valgrind)
   if(VALGRIND_EXE AND NOT CMAKE_CROSSCOMPILING)
@@ -285,6 +177,12 @@ if(BUILD_TESTING)
                      $<TARGET_FILE:driver> ${CORPUS_TESTS})
   endif()
 
+  foreach(TEST ${CORPUS_TESTS})
+    get_filename_component(hash ${TEST} NAME)
+    add_test(NAME conflict_set_blackbox_${hash} COMMAND driver ${TEST})
+    add_test(NAME skip_list_${hash} COMMAND driver_skip_list ${TEST})
+  endforeach()
+
   # api smoke tests
 
   # c90
@@ -305,45 +203,31 @@ if(BUILD_TESTING)
                         PROPERTIES CXX_STANDARD_REQUIRED ON)
   add_test(NAME conflict_set_cxx_api_test COMMAND conflict_set_cxx_api_test)
 
-  # symbol visibility tests
+  if(NOT APPLE AND NOT CMAKE_BUILD_TYPE STREQUAL Debug)
-  if(NOT WASM AND NOT CMAKE_BUILD_TYPE STREQUAL Debug)
-    if(APPLE)
-      set(symbol_exports ${CMAKE_SOURCE_DIR}/apple-symbol-exports.txt)
-      set(symbol_imports ${CMAKE_SOURCE_DIR}/apple-symbol-imports.txt)
-    else()
-      set(symbol_exports ${CMAKE_SOURCE_DIR}/symbol-exports.txt)
-      if(CMAKE_SYSTEM_PROCESSOR STREQUAL aarch64)
-        set(symbol_imports ${CMAKE_SOURCE_DIR}/aarch64-symbol-imports.txt)
-      else()
-        set(symbol_imports ${CMAKE_SOURCE_DIR}/symbol-imports.txt)
-      endif()
-    endif()
     add_test(
       NAME conflict_set_shared_symbols
-      COMMAND
+      COMMAND ${CMAKE_SOURCE_DIR}/test_symbols.sh
-        ${CMAKE_SOURCE_DIR}/test_symbols.sh $<TARGET_FILE:${PROJECT_NAME}>
+              $<TARGET_FILE:${PROJECT_NAME}> ${CMAKE_SOURCE_DIR}/symbols.txt)
-        ${symbol_exports} ${symbol_imports})
     add_test(
       NAME conflict_set_static_symbols
       COMMAND
         ${CMAKE_SOURCE_DIR}/test_symbols.sh
-        $<TARGET_FILE:${PROJECT_NAME}-static> ${symbol_exports}
+        $<TARGET_FILE:${PROJECT_NAME}_static> ${CMAKE_SOURCE_DIR}/symbols.txt)
-        ${symbol_imports})
   endif()
 
   # bench
+
   add_executable(conflict_set_bench Bench.cpp)
   target_link_libraries(conflict_set_bench PRIVATE ${PROJECT_NAME})
   set_target_properties(conflict_set_bench PROPERTIES SKIP_BUILD_RPATH ON)
+
   add_executable(real_data_bench RealDataBench.cpp)
   target_link_libraries(real_data_bench PRIVATE ${PROJECT_NAME})
   set_target_properties(real_data_bench PROPERTIES SKIP_BUILD_RPATH ON)
 endif()
-
 # packaging
 
 set(CPACK_PACKAGE_CONTACT andrew@weaselab.dev)
-set(CMAKE_INSTALL_DEFAULT_COMPONENT_NAME all)
 
 set(CPACK_PACKAGE_VENDOR "Weaselab")
 set(CPACK_RESOURCE_FILE_LICENSE "${CMAKE_CURRENT_SOURCE_DIR}/LICENSE")
@@ -353,30 +237,6 @@ set(CPACK_RESOURCE_FILE_README "${CMAKE_CURRENT_SOURCE_DIR}/README.md")
 set(CPACK_RPM_PACKAGE_ARCHITECTURE ${CMAKE_SYSTEM_PROCESSOR})
 set(CPACK_RPM_SPEC_INSTALL_POST "/bin/true") # avoid stripping
 set(CPACK_RPM_PACKAGE_LICENSE "Apache 2.0")
-set(CPACK_RPM_FILE_NAME RPM-DEFAULT)
-
-# deb
-set(CPACK_DEBIAN_FILE_NAME DEB-DEFAULT)
-# see *-imports.txt - dependency versions need to be synced with symbol versions
-if(CMAKE_SYSTEM_PROCESSOR STREQUAL aarch64)
-  set(CPACK_DEBIAN_PACKAGE_DEPENDS "libc6 (>= 2.17)")
-else()
-  set(CPACK_DEBIAN_PACKAGE_DEPENDS "libc6 (>= 2.14)")
-endif()
-
-# macos
-set(CMAKE_OSX_DEPLOYMENT_TARGET 11.0)
-if(APPLE)
-  find_program(PANDOC_EXE pandoc)
-  if(PANDOC_EXE)
-    execute_process(COMMAND ${PANDOC_EXE} ${CMAKE_SOURCE_DIR}/README.md -o
-                            ${CMAKE_BINARY_DIR}/README.txt)
-    set(CPACK_RESOURCE_FILE_README ${CMAKE_BINARY_DIR}/README.txt)
-  endif()
-  configure_file(${CMAKE_SOURCE_DIR}/LICENSE ${CMAKE_BINARY_DIR}/LICENSE.txt
-                 COPYONLY)
-  set(CPACK_RESOURCE_FILE_LICENSE ${CMAKE_BINARY_DIR}/LICENSE.txt)
-endif()
 
 include(CPack)
 
@@ -388,7 +248,7 @@ target_include_directories(
          $<INSTALL_INTERFACE:${CMAKE_INSTALL_INCLUDEDIR}/${PROJECT_NAME}>)
 
 target_include_directories(
-  ${PROJECT_NAME}-static
+  ${PROJECT_NAME}_static
   PUBLIC $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/include>
          $<INSTALL_INTERFACE:${CMAKE_INSTALL_INCLUDEDIR}/${PROJECT_NAME}>)
 
@@ -397,16 +257,13 @@ set_target_properties(
                              SOVERSION ${PROJECT_VERSION_MAJOR})
 
 install(
-  TARGETS ${PROJECT_NAME} ${PROJECT_NAME}-static
+  TARGETS ${PROJECT_NAME} ${PROJECT_NAME}_static
-  EXPORT ${PROJECT_NAME}Config
+  EXPORT ConflictSetConfig
   ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR}
   LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
   RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR})
-
 install(DIRECTORY include/
         DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/${PROJECT_NAME})
 
-install(EXPORT ${PROJECT_NAME}Config
+install(EXPORT ConflictSetConfig
-        DESTINATION ${CMAKE_INSTALL_DATAROOTDIR}/${PROJECT_NAME}/cmake)
+        DESTINATION ${CMAKE_INSTALL_DATAROOTDIR}/ConflictSet/cmake)
-
-cpack_add_component(all)

FuzzTestDriver.cpp

@@ -2,25 +2,15 @@
 #include <cstdint>
 #include <fstream>
 #include <sstream>
-#include <thread>
 
 extern "C" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size);
 
 int main(int argc, char **argv) {
-  auto doTest = [&]() {
+  for (int i = 1; i < argc; ++i) {
-    for (int i = 1; i < argc; ++i) {
+    std::ifstream t(argv[i], std::ios::binary);
-      std::ifstream t(argv[i], std::ios::binary);
+    std::stringstream buffer;
-      std::stringstream buffer;
+    buffer << t.rdbuf();
-      buffer << t.rdbuf();
+    auto str = buffer.str();
-      auto str = buffer.str();
+    LLVMFuzzerTestOneInput((const uint8_t *)str.data(), str.size());
-      LLVMFuzzerTestOneInput((const uint8_t *)str.data(), str.size());
+  }
-    }
-  };
-#ifdef THREAD_TEST
-  std::thread thread2{doTest};
-#endif
-  doTest();
-#ifdef THREAD_TEST
-  thread2.join();
-#endif
 }

HashTable.cpp

@@ -96,15 +96,13 @@ void ConflictSet::setOldestVersion(int64_t oldestVersion) {
   return impl->setOldestVersion(oldestVersion);
 }
 
-int64_t ConflictSet::getBytes() const { return -1; }
-
 ConflictSet::ConflictSet(int64_t oldestVersion)
-    : impl(new(safe_malloc(sizeof(Impl))) Impl{oldestVersion}) {}
+    : impl(new (safe_malloc(sizeof(Impl))) Impl{oldestVersion}) {}
 
 ConflictSet::~ConflictSet() {
   if (impl) {
     impl->~Impl();
-    safe_free(impl, sizeof(Impl));
+    free(impl);
   }
 }
 
@@ -144,11 +142,6 @@ ConflictSet_create(int64_t oldestVersion) {
 __attribute__((__visibility__("default"))) void ConflictSet_destroy(void *cs) {
   using Impl = ConflictSet::Impl;
   ((Impl *)cs)->~Impl();
-  safe_free(cs, sizeof(Impl));
+  free(cs);
-}
-__attribute__((__visibility__("default"))) int64_t
-ConflictSet_getBytes(void *cs) {
-  using Impl = ConflictSet::Impl;
-  return -1;
 }
 }

Internal.h

@@ -2,8 +2,6 @@
 
 #include "ConflictSet.h"
 
-using namespace weaselab;
-
 #include <bit>
 #include <cassert>
 #include <compare>
@@ -12,12 +10,10 @@ using namespace weaselab;
 #include <cstdlib>
 #include <cstring>
 #include <inttypes.h>
-#include <latch>
 #include <map>
 #include <set>
 #include <span>
 #include <string>
-#include <thread>
 #include <unordered_set>
 #include <utility>
 #include <vector>
@@ -39,76 +35,16 @@ operator<=>(const std::span<const uint8_t> &lhs,
   return lhs.size() <=> rhs.size();
 }
 
-[[nodiscard]] inline auto operator<=>(const std::span<const uint8_t> &lhs,
-                                      const ConflictSet::Key &rhs) noexcept {
-  int cl = std::min<int>(lhs.size(), rhs.len);
-  if (cl > 0) {
-    if (auto c = memcmp(lhs.data(), rhs.p, cl) <=> 0; c != 0) {
-      return c;
-    }
-  }
-  return lhs.size() <=> size_t(rhs.len);
-}
-
 // This header contains code that we want to reuse outside of ConflictSet.cpp or
 // want to exclude from coverage since it's only testing related.
 
 // GCOVR_EXCL_START
 
-#if SHOW_MEMORY
-inline int64_t mallocBytes = 0;
-inline int64_t peakMallocBytes = 0;
-#endif
-
-inline thread_local int64_t mallocBytesDelta = 0;
-
-#ifndef NDEBUG
-constexpr auto kMallocHeaderSize = 16;
-#endif
-
-// malloc that aborts on OOM and thus always returns a non-null pointer. Must be
-// paired with `safe_free`.
 __attribute__((always_inline)) inline void *safe_malloc(size_t s) {
-  mallocBytesDelta += s;
+  if (void *p = malloc(s)) {
-#if SHOW_MEMORY
+    return p;
-  mallocBytes += s;
-  if (mallocBytes > peakMallocBytes) {
-    peakMallocBytes = mallocBytes;
   }
-#endif
+  abort();
-  void *p = malloc(s
-#ifndef NDEBUG
-                   + kMallocHeaderSize
-#endif
-  );
-  if (p == nullptr) {
-    abort();
-  }
-#ifndef NDEBUG
-  memcpy(p, &s, sizeof(s));
-  (char *&)p += kMallocHeaderSize;
-#endif
-  return p;
-}
-
-// Must be paired with `safe_malloc`.
-//
-// There's nothing safer about this than free. Only called safe_free for
-// symmetry with safe_malloc.
-__attribute__((always_inline)) inline void safe_free(void *p, size_t s) {
-  mallocBytesDelta -= s;
-#if SHOW_MEMORY
-  mallocBytes -= s;
-  free(p);
-#else
-#ifndef NDEBUG
-  (char *&)p -= kMallocHeaderSize;
-  size_t expected;
-  memcpy(&expected, p, sizeof(expected));
-  assert(s == expected);
-#endif
-  free(p);
-#endif
 }
 
 // ==================== BEGIN ARENA IMPL ====================
@@ -154,7 +90,6 @@ inline void *operator new[](size_t size, std::align_val_t align,
 
 /// align must be a power of two
 template <class T> T *align_up(T *t, size_t align) {
-  assert(std::popcount(align) == 1);
   auto unaligned = uintptr_t(t);
   auto aligned = (unaligned + align - 1) & ~(align - 1);
   return reinterpret_cast<T *>(reinterpret_cast<char *>(t) + aligned -
@@ -163,7 +98,6 @@ template <class T> T *align_up(T *t, size_t align) {
 
 /// align must be a power of two
 constexpr inline int align_up(uint32_t unaligned, uint32_t align) {
-  assert(std::popcount(align) == 1);
   return (unaligned + align - 1) & ~(align - 1);
 }
 
@@ -179,10 +113,12 @@ struct Arena::ArenaImpl {
   uint8_t *begin() { return reinterpret_cast<uint8_t *>(this + 1); }
 };
 
+static_assert(sizeof(Arena::ArenaImpl) == 16);
+static_assert(alignof(Arena::ArenaImpl) == 8);
+
 inline Arena::Arena(int initialSize) : impl(nullptr) {
   if (initialSize > 0) {
-    auto allocationSize =
+    auto allocationSize = align_up(initialSize + sizeof(ArenaImpl), 16);
-        align_up(initialSize + sizeof(ArenaImpl), alignof(ArenaImpl));
     impl = (Arena::ArenaImpl *)safe_malloc(allocationSize);
     impl->prev = nullptr;
     impl->capacity = allocationSize - sizeof(ArenaImpl);
@@ -193,7 +129,7 @@ inline Arena::Arena(int initialSize) : impl(nullptr) {
 inline void onDestroy(Arena::ArenaImpl *impl) {
   while (impl) {
     auto *prev = impl->prev;
-    safe_free(impl, sizeof(Arena::ArenaImpl) + impl->capacity);
+    free(impl);
     impl = prev;
   }
 }
@@ -218,7 +154,7 @@ inline void *operator new(size_t size, std::align_val_t align, Arena &arena) {
                           (arena.impl ? std::max<int>(sizeof(Arena::ArenaImpl),
                                                       arena.impl->capacity * 2)
                                       : 0)),
-        alignof(Arena::ArenaImpl));
+        16);
     auto *impl = (Arena::ArenaImpl *)safe_malloc(allocationSize);
     impl->prev = arena.impl;
     impl->capacity = allocationSize - sizeof(Arena::ArenaImpl);
@@ -413,6 +349,34 @@ inline uint32_t Arbitrary::bounded(uint32_t s) {
 
 // ==================== END ARBITRARY IMPL ====================
 
+// ==================== BEGIN UTILITIES IMPL ====================
+
+// Call Stepwise::step for each element of remaining until it returns true.
+// Applies a permutation to `remaining` as a side effect.
+template <class Stepwise> void runInterleaved(std::span<Stepwise> remaining) {
+  while (remaining.size() > 0) {
+    for (int i = 0; i < int(remaining.size());) {
+      bool done = remaining[i].step();
+      if (done) {
+        if (i != int(remaining.size()) - 1) {
+          using std::swap;
+          swap(remaining[i], remaining.back());
+        }
+        remaining = remaining.subspan(0, remaining.size() - 1);
+      } else {
+        ++i;
+      }
+    }
+  }
+};
+
+template <class Stepwise> void runSequential(std::span<Stepwise> remaining) {
+  for (auto &r : remaining) {
+    while (!r.step()) {
+    }
+  }
+}
+
 struct ReferenceImpl {
   explicit ReferenceImpl(int64_t oldestVersion) : oldestVersion(oldestVersion) {
     writeVersionMap[""] = oldestVersion;
@@ -506,18 +470,6 @@ inline std::string printable(std::span<const uint8_t> key) {
   return printable(std::string_view((const char *)key.data(), key.size()));
 }
 
-inline const char *resultToStr(ConflictSet::Result r) {
-  switch (r) {
-  case ConflictSet::Commit:
-    return "commit";
-  case ConflictSet::Conflict:
-    return "conflict";
-  case ConflictSet::TooOld:
-    return "too old";
-  }
-  abort();
-}
-
 namespace {
 
 template <class ConflictSetImpl> struct TestDriver {
@@ -530,12 +482,21 @@ template <class ConflictSetImpl> struct TestDriver {
   ConflictSetImpl cs{oldestVersion};
   ReferenceImpl refImpl{oldestVersion};
 
-  constexpr static auto kMaxKeySuffixLen = 8;
+  constexpr static auto kMaxKeyLen = 64;
 
   bool ok = true;
 
-  const int prefixLen = arbitrary.bounded(512);
+  static const char *resultToStr(ConflictSet::Result r) {
-  const int prefixByte = arbitrary.randT<uint8_t>();
+    switch (r) {
+    case ConflictSet::Commit:
+      return "commit";
+    case ConflictSet::Conflict:
+      return "conflict";
+    case ConflictSet::TooOld:
+      return "too old";
+    }
+    abort();
+  }
 
   // Call until it returns true, for "done". Check internal invariants etc
   // between calls to next.
@@ -555,10 +516,9 @@ template <class ConflictSetImpl> struct TestDriver {
         if (!arbitrary.hasEntropy()) {
           return true;
         }
-        int keyLen = prefixLen + arbitrary.bounded(kMaxKeySuffixLen);
+        int keyLen = arbitrary.bounded(kMaxKeyLen);
         auto *begin = new (arena) uint8_t[keyLen];
-        memset(begin, prefixByte, prefixLen);
+        arbitrary.randomBytes(begin, keyLen);
-        arbitrary.randomBytes(begin + prefixLen, keyLen - prefixLen);
         keys.insert(std::string_view((const char *)begin, keyLen));
       }
 
@@ -623,10 +583,9 @@ template <class ConflictSetImpl> struct TestDriver {
         if (!arbitrary.hasEntropy()) {
           return true;
         }
-        int keyLen = prefixLen + arbitrary.bounded(kMaxKeySuffixLen);
+        int keyLen = arbitrary.bounded(kMaxKeyLen);
         auto *begin = new (arena) uint8_t[keyLen];
-        memset(begin, prefixByte, prefixLen);
+        arbitrary.randomBytes(begin, keyLen);
-        arbitrary.randomBytes(begin + prefixLen, keyLen - prefixLen);
         keys.insert(std::string_view((const char *)begin, keyLen));
       }
 
@@ -673,60 +632,32 @@ template <class ConflictSetImpl> struct TestDriver {
       auto *results2 =
           new (arena) ConflictSet::Result[numPointReads + numRangeReads];
 
-#ifdef THREAD_TEST
-      auto *results3 =
-          new (arena) ConflictSet::Result[numPointReads + numRangeReads];
-      std::latch ready{1};
-      std::thread thread2{[&]() {
-        ready.count_down();
-        cs.check(reads, results3, numPointReads + numRangeReads);
-      }};
-      ready.wait();
-#endif
-
       CALLGRIND_START_INSTRUMENTATION;
       cs.check(reads, results1, numPointReads + numRangeReads);
       CALLGRIND_STOP_INSTRUMENTATION;
 
       refImpl.check(reads, results2, numPointReads + numRangeReads);
-
+      for (int i = 0; i < numPointReads + numRangeReads; ++i) {
-      auto compareResults = [reads](ConflictSet::Result *results1,
+        if (results1[i] != results2[i]) {
-                                    ConflictSet::Result *results2, int count) {
+          if (reads[i].end.len == 0) {
-        for (int i = 0; i < count; ++i) {
+            fprintf(stderr,
-          if (results1[i] != results2[i]) {
+                    "Expected %s, got %s for read of {%s} at version %" PRId64
-            if (reads[i].end.len == 0) {
+                    "\n",
-              fprintf(stderr,
+                    resultToStr(results2[i]), resultToStr(results1[i]),
-                      "Expected %s, got %s for read of {%s} at version %" PRId64
+                    printable(reads[i].begin).c_str(), reads[i].readVersion);
-                      "\n",
+          } else {
-                      resultToStr(results2[i]), resultToStr(results1[i]),
+            fprintf(
-                      printable(reads[i].begin).c_str(), reads[i].readVersion);
+                stderr,
-            } else {
+                "Expected %s, got %s for read of [%s, %s) at version %" PRId64
-              fprintf(
+                "\n",
-                  stderr,
+                resultToStr(results2[i]), resultToStr(results1[i]),
-                  "Expected %s, got %s for read of [%s, %s) at version %" PRId64
+                printable(reads[i].begin).c_str(),
-                  "\n",
+                printable(reads[i].end).c_str(), reads[i].readVersion);
-                  resultToStr(results2[i]), resultToStr(results1[i]),
-                  printable(reads[i].begin).c_str(),
-                  printable(reads[i].end).c_str(), reads[i].readVersion);
-            }
-            return false;
           }
+          ok = false;
+          return true;
         }
-        return true;
-      };
-
-      if (!compareResults(results1, results2, numPointReads + numRangeReads)) {
-        ok = false;
-        return true;
       }
-
-#ifdef THREAD_TEST
-      thread2.join();
-      if (!compareResults(results3, results2, numPointReads + numRangeReads)) {
-        ok = false;
-        return true;
-      }
-#endif
     }
     return false;
   }

Jenkinsfile

@@ -36,29 +36,6 @@ pipeline {
                 sh 'pre-commit run --all-files --show-diff-on-failure'
             }
         }
-        stage('Clang') {
-            agent {
-                dockerfile {
-                    args '-v /home/jenkins/ccache:/ccache'
-                    reuseNode true
-                }
-            }
-            steps {
-                CleanBuildAndTest("")
-                recordIssues(tools: [clang()])
-            }
-        }
-        stage('SIMD fallback') {
-            agent {
-                dockerfile {
-                    args '-v /home/jenkins/ccache:/ccache'
-                    reuseNode true
-                }
-            }
-            steps {
-                CleanBuildAndTest("-DUSE_SIMD_FALLBACK=ON")
-            }
-        }
         stage('Release [gcc]') {
             agent {
                 dockerfile {
@@ -67,7 +44,7 @@ pipeline {
                 }
             }
             steps {
-                CleanBuildAndTest("-DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++ -DCMAKE_CXX_FLAGS=-DNVALGRIND")
+                CleanBuildAndTest("-DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++ -DCMAKE_BUILD_TYPE=Release")
                 recordIssues(tools: [gcc()])
                 sh '''
                 cd build
@@ -89,7 +66,7 @@ pipeline {
                 }
             }
             steps {
-                CleanBuildAndTest("-DCMAKE_TOOLCHAIN_FILE=../aarch64-toolchain.cmake -DCMAKE_CXX_FLAGS=-DNVALGRIND")
+                CleanBuildAndTest("-DCMAKE_TOOLCHAIN_FILE=../aarch64-toolchain.cmake")
                 sh '''
                 cd build
                 cpack -G DEB
@@ -108,12 +85,9 @@ pipeline {
             steps {
                 CleanBuildAndTest("-DCMAKE_C_COMPILER=gcc -DCMAKE_CXX_COMPILER=g++ -DCMAKE_C_FLAGS=--coverage -DCMAKE_CXX_FLAGS=--coverage -DCMAKE_BUILD_TYPE=Debug")
                 sh '''
-                gcovr -f ConflictSet.cpp --cobertura > build/coverage.xml
+                gcovr --exclude '.*third_party.*' --cobertura > build/coverage.xml
                 '''
                 cobertura autoUpdateHealth: false, autoUpdateStability: false, coberturaReportFile: 'build/coverage.xml', conditionalCoverageTargets: '70, 0, 0', failUnhealthy: false, failUnstable: false, lineCoverageTargets: '80, 0, 0', maxNumberOfBuilds: 0, methodCoverageTargets: '80, 0, 0', onlyStable: false, sourceEncoding: 'ASCII', zoomCoverageChart: false
-                sh '''
-                gcovr -f ConflictSet.cpp --fail-under-line 100 > /dev/null
-                '''
             }
         }
     }

README.md

@@ -2,83 +2,81 @@ A data structure for optimistic concurrency control on ranges of bitwise-lexicog
 
 Intended to replace FoundationDB's skip list.
 
-Hardware for all benchmarks is a mac m1 2020.
-
 # FoundationDB's benchmark
 
 ## Skip list
 
 ```
-New conflict set: 1.957 sec
+New conflict set: 2.404 sec
-                  0.639 Mtransactions/sec
+                  0.520 Mtransactions/sec
-                  2.555 Mkeys/sec
+                  2.080 Mkeys/sec
-Detect only:      1.845 sec
+Detect only:      2.266 sec
-                  0.678 Mtransactions/sec
+                  0.552 Mtransactions/sec
-                  2.710 Mkeys/sec
+                  2.207 Mkeys/sec
-Skiplist only:    1.263 sec
+Skiplist only:    1.594 sec
-                  0.990 Mtransactions/sec
+                  0.784 Mtransactions/sec
-                  3.960 Mkeys/sec
+                  3.137 Mkeys/sec
 Performance counters:
-               Build: 0.0546
+               Build: 0.071
-                 Add: 0.0563
+                 Add: 0.0641
-              Detect: 1.84
+              Detect: 2.27
-              D.Sort: 0.412
+              D.Sort: 0.44
-           D.Combine: 0.0141
+           D.Combine: 0.018
-         D.CheckRead: 0.671
+         D.CheckRead: 0.855
-   D.CheckIntraBatch: 0.0068
+   D.CheckIntraBatch: 0.00903
-        D.MergeWrite: 0.592
+        D.MergeWrite: 0.739
-      D.RemoveBefore: 0.146
+      D.RemoveBefore: 0.201
 ```
 
 ## Radix tree (this implementation)
 
 ```
-New conflict set: 1.366 sec
+New conflict set: 1.743 sec
-                  0.915 Mtransactions/sec
+                  0.717 Mtransactions/sec
-                  3.660 Mkeys/sec
+                  2.869 Mkeys/sec
-Detect only:      1.248 sec
+Detect only:      1.611 sec
-                  1.002 Mtransactions/sec
+                  0.776 Mtransactions/sec
-                  4.007 Mkeys/sec
+                  3.103 Mkeys/sec
-Skiplist only:    0.573 sec
+Skiplist only:    0.919 sec
-                  2.182 Mtransactions/sec
+                  1.360 Mtransactions/sec
-                  8.730 Mkeys/sec
+                  5.440 Mkeys/sec
 Performance counters:
-               Build: 0.0594
+               Build: 0.0657
-                 Add: 0.0572
+                 Add: 0.0628
-              Detect: 1.25
+              Detect: 1.61
-              D.Sort: 0.418
+              D.Sort: 0.442
-           D.Combine: 0.0149
+           D.Combine: 0.0178
-         D.CheckRead: 0.232
+         D.CheckRead: 0.395
-   D.CheckIntraBatch: 0.0067
+   D.CheckIntraBatch: 0.00776
-        D.MergeWrite: 0.341
+        D.MergeWrite: 0.524
-      D.RemoveBefore: 0.232
+      D.RemoveBefore: 0.221
 ```
 
 # Our benchmark
 
 ## Skip list
 
-|  ns/op |         op/s | err% | total | benchmark                           |
+|               ns/op |                op/s |    err% |     total | benchmark
-| -----: | -----------: | ---: | ----: | :---------------------------------- |
+|--------------------:|--------------------:|--------:|----------:|:----------
-| 246.99 | 4,048,700.59 | 0.2% |  0.01 | `point reads`                       |
+|              270.07 |        3,702,706.03 |    0.4% |      0.01 | `point reads`
-| 260.16 | 3,843,784.65 | 0.1% |  0.01 | `prefix reads`                      |
+|              285.76 |        3,499,437.03 |    1.5% |      0.01 | `prefix reads`
-| 493.35 | 2,026,953.19 | 0.1% |  0.01 | `range reads`                       |
+|              532.54 |        1,877,794.90 |    0.7% |      0.01 | `range reads`
-| 462.05 | 2,164,289.23 | 0.6% |  0.01 | `point writes`                      |
+|              528.50 |        1,892,132.94 |    0.7% |      0.01 | `point writes`
-| 448.19 | 2,231,205.25 | 0.9% |  0.01 | `prefix writes`                     |
+|              516.53 |        1,935,978.22 |    0.9% |      0.01 | `prefix writes`
-| 255.83 | 3,908,845.72 | 1.5% |  0.02 | `range writes`                      |
+|              303.34 |        3,296,630.84 |    3.6% |      0.05 | `range writes`
-| 582.63 | 1,716,349.02 | 1.3% |  0.01 | `monotonic increasing point writes` |
+|              502.88 |        1,988,553.24 |    2.0% |      0.01 | `monotonic increasing point writes`
 
 ## Radix tree (this implementation)
 
-|  ns/op |          op/s | err% | total | benchmark                           |
+|               ns/op |                op/s |    err% |     total | benchmark
-| -----: | ------------: | ---: | ----: | :---------------------------------- |
+|--------------------:|--------------------:|--------:|----------:|:----------
-|  19.42 | 51,483,206.67 | 0.3% |  0.01 | `point reads`                       |
+|               14.52 |       68,850,842.99 |    1.2% |      0.01 | `point reads`
-|  58.43 | 17,115,612.57 | 0.1% |  0.01 | `prefix reads`                      |
+|               60.89 |       16,422,538.22 |    1.5% |      0.01 | `prefix reads`
-| 216.09 |  4,627,766.60 | 0.2% |  0.01 | `range reads`                       |
+|              226.89 |        4,407,362.98 |    0.5% |      0.01 | `range reads`
-|  28.35 | 35,267,567.72 | 0.2% |  0.01 | `point writes`                      |
+|               22.99 |       43,498,198.49 |    0.2% |      0.01 | `point writes`
-|  43.43 | 23,026,226.17 | 0.2% |  0.01 | `prefix writes`                     |
+|               50.51 |       19,799,864.54 |    1.0% |      0.01 | `prefix writes`
-|  50.00 | 20,000,000.00 | 0.0% |  0.01 | `range writes`                      |
+|               82.50 |       12,121,212.12 |    2.6% |      0.03 | `range writes`
-|  92.38 | 10,824,863.69 | 4.1% |  0.01 | `monotonic increasing point writes` |
+|              119.94 |        8,337,354.54 |    2.1% |      0.01 | `monotonic increasing point writes`
 
 # "Real data" test
 
@@ -87,13 +85,13 @@ Point queries only, best of three runs. Gc ratio is the ratio of time spent doin
 ## skip list
 
 ```
-Check: 11.3385 seconds, 329.718 MB/s, Add: 5.35612 seconds, 131.072 MB/s, Gc ratio: 45.7173%
+Check: 12.7863 seconds, 292.384 MB/s, Add: 19.8276 seconds, 35.4071 MB/s, Gc ratio: 23.5314%
 ```
 
 ## radix tree
 
 ```
-Check: 2.48583 seconds, 1503.93 MB/s, Add: 2.12768 seconds, 329.954 MB/s, Gc ratio: 41.7943%
+Check: 3.60187 seconds, 1037.94 MB/s, Add: 3.03958 seconds, 230.966 MB/s, Gc ratio: 52.3876%
 ```
 
 ## hash table
@@ -101,5 +99,5 @@ Check: 2.48583 seconds, 1503.93 MB/s, Add: 2.12768 seconds, 329.954 MB/s, Gc rat
 (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: 1.83386 seconds, 2038.6 MB/s, Add: 0.601411 seconds, 1167.32 MB/s, Gc ratio: 48.9776%
+Check: 2.15925 seconds, 1731.4 MB/s, Add: 1.08519 seconds, 646.926 MB/s, Gc ratio: 52.1526%
 ```

RealDataBench.cpp

@@ -1,18 +1,13 @@
 #include <ConflictSet.h>
 
-#include <cerrno>
 #include <chrono>
-#include <cstdio>
 #include <cstring>
 #include <fcntl.h>
-#include <string_view>
 #include <sys/mman.h>
 #include <sys/stat.h>
 #include <unistd.h>
 #include <vector>
 
-using namespace weaselab;
-
 double now() {
   return std::chrono::duration_cast<std::chrono::nanoseconds>(
              std::chrono::steady_clock::now().time_since_epoch())
@@ -33,7 +28,7 @@ constexpr inline size_t rightAlign(size_t offset, size_t alignment) {
 
 int main(int argc, const char **argv) {
   // Use with this dataset https://snap.stanford.edu/data/memetracker9.html
-  // Preprocess the files with `sed -i'' '/^Q/d'`
+  // Preprocess the files with `sed -i '' '/^Q/d'`
 
   double checkTime = 0;
   double addTime = 0;
@@ -45,8 +40,6 @@ int main(int argc, const char **argv) {
   int64_t version = 0;
   double timer = 0;
 
-  int64_t peakMemory = 0;
-
   for (int i = 1; i < argc; ++i) {
     int fd = open(argv[i], O_RDONLY);
     struct stat st;
@@ -116,10 +109,6 @@ int main(int argc, const char **argv) {
         write = {};
         reads.clear();
 
-        if (cs.getBytes() > peakMemory) {
-          peakMemory = cs.getBytes();
-        }
-
         timer = now();
         cs.setOldestVersion(version - 10000);
         gcTime += now() - timer;
@@ -129,9 +118,8 @@ int main(int argc, const char **argv) {
     close(fd);
   }
 
-  printf("Check: %g seconds, %g MB/s, Add: %g seconds, %g MB/s, Gc ratio: "
+  printf(
-         "%g%%, Peak idle memory: %g\n",
+      "Check: %g seconds, %g MB/s, Add: %g seconds, %g MB/s, Gc ratio: %g%%\n",
-         checkTime, checkBytes / checkTime * 1e-6, addTime,
+      checkTime, checkBytes / checkTime * 1e-6, addTime,
-         addBytes / addTime * 1e-6, gcTime / (gcTime + addTime) * 1e2,
+      addBytes / addTime * 1e-6, gcTime / (gcTime + addTime) * 1e2);
-         double(peakMemory));
 }

SkipList.cpp

@@ -16,8 +16,6 @@
  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  * See the License for the specific language governing permissions and
  * limitations under the License.
- *
- * This source code is modified to compile outside of FoundationDB
  */
 
 #include "ConflictSet.h"
@@ -151,7 +149,7 @@ private:
       setMaxVersion(level, v);
     }
 
-    void destroy() { safe_free(this, getNodeSize()); }
+    void destroy() { free(this); }
 
   private:
     int getNodeSize() const {
@@ -270,21 +268,13 @@ public:
   }
 
   explicit SkipList(Version version = 0) {
-#if DEBUG_VERBOSE
-    fprintf(stderr, "skip_list: create\n");
-#endif
     header = Node::create(StringRef(), MaxLevels - 1);
     for (int l = 0; l < MaxLevels; l++) {
       header->setNext(l, nullptr);
       header->setMaxVersion(l, version);
     }
   }
-  ~SkipList() {
+  ~SkipList() { destroy(); }
-#if DEBUG_VERBOSE
-    fprintf(stderr, "skip_list: destroy\n");
-#endif
-    destroy();
-  }
   SkipList(SkipList &&other) noexcept : header(other.header) {
     other.header = nullptr;
   }
@@ -613,15 +603,11 @@ struct __attribute__((visibility("hidden"))) ConflictSet::Impl {
     for (int s = stripes - 1; s >= 0; s--) {
       for (int i = 0; i * 2 < ss; ++i) {
         const auto &w = writes[s * stripeSize / 2 + i];
-#if DEBUG_VERBOSE
-        printf("Write begin: %s\n", printable(w.begin).c_str());
-        fflush(stdout);
-#endif
         values[i * 2] = {w.begin.p, size_t(w.begin.len)};
         values[i * 2 + 1] = w.end.len > 0
                                 ? StringRef{w.end.p, size_t(w.end.len)}
                                 : keyAfter(arena, values[i * 2]);
-        keyUpdates += 3;
+        keyUpdates += 2;
       }
       skipList.find(values, fingers, temp, ss);
       skipList.addConflictRanges(fingers, ss / 2, writeVersion);
@@ -635,16 +621,14 @@ struct __attribute__((visibility("hidden"))) ConflictSet::Impl {
     int temp;
     std::span<const uint8_t> key = removalKey;
     skipList.find(&key, &finger, &temp, 1);
-    skipList.removeBefore(oldestVersion, finger, std::exchange(keyUpdates, 10));
+    skipList.removeBefore(oldestVersion, finger, std::exchange(keyUpdates, 0));
     removalArena = Arena();
     removalKey = copyToArena(
         removalArena, {finger.getValue().data(), finger.getValue().size()});
   }
 
-  int64_t totalBytes = 0;
-
 private:
-  int64_t keyUpdates = 10;
+  int64_t keyUpdates = 0;
   Arena removalArena;
   std::span<const uint8_t> removalKey;
   int64_t oldestVersion;
@@ -653,44 +637,25 @@ private:
 
 void ConflictSet::check(const ReadRange *reads, Result *results,
                         int count) const {
-  impl->check(reads, results, count);
+  return impl->check(reads, results, count);
 }
 
 void ConflictSet::addWrites(const WriteRange *writes, int count,
                             int64_t writeVersion) {
-  mallocBytesDelta = 0;
+  return impl->addWrites(writes, count, writeVersion);
-  impl->addWrites(writes, count, writeVersion);
-  impl->totalBytes += mallocBytesDelta;
-#if SHOW_MEMORY
-  if (impl->totalBytes != mallocBytes) {
-    abort();
-  }
-#endif
 }
 
 void ConflictSet::setOldestVersion(int64_t oldestVersion) {
-  mallocBytesDelta = 0;
+  return impl->setOldestVersion(oldestVersion);
-  impl->setOldestVersion(oldestVersion);
-  impl->totalBytes += mallocBytesDelta;
-#if SHOW_MEMORY
-  if (impl->totalBytes != mallocBytes) {
-    abort();
-  }
-#endif
 }
 
-int64_t ConflictSet::getBytes() const { return impl->totalBytes; }
-
 ConflictSet::ConflictSet(int64_t oldestVersion)
-    : impl((mallocBytesDelta = 0,
+    : impl(new (safe_malloc(sizeof(Impl))) Impl{oldestVersion}) {}
-            new(safe_malloc(sizeof(Impl))) Impl{oldestVersion})) {
-  impl->totalBytes += mallocBytesDelta;
-}
 
 ConflictSet::~ConflictSet() {
   if (impl) {
     impl->~Impl();
-    safe_free(impl, sizeof(Impl));
+    free(impl);
   }
 }
 
@@ -716,54 +681,20 @@ ConflictSet_check(void *cs, const ConflictSet_ReadRange *reads,
 __attribute__((__visibility__("default"))) void
 ConflictSet_addWrites(void *cs, const ConflictSet_WriteRange *writes, int count,
                       int64_t writeVersion) {
-  auto *impl = (ConflictSet::Impl *)cs;
+  ((ConflictSet::Impl *)cs)->addWrites(writes, count, writeVersion);
-  mallocBytesDelta = 0;
-  impl->addWrites(writes, count, writeVersion);
-  impl->totalBytes += mallocBytesDelta;
-#if SHOW_MEMORY
-  if (impl->totalBytes != mallocBytes) {
-    abort();
-  }
-#endif
 }
 __attribute__((__visibility__("default"))) void
 ConflictSet_setOldestVersion(void *cs, int64_t oldestVersion) {
-  auto *impl = (ConflictSet::Impl *)cs;
+  ((ConflictSet::Impl *)cs)->setOldestVersion(oldestVersion);
-  mallocBytesDelta = 0;
-  impl->setOldestVersion(oldestVersion);
-  impl->totalBytes += mallocBytesDelta;
-#if SHOW_MEMORY
-  if (impl->totalBytes != mallocBytes) {
-    abort();
-  }
-#endif
 }
 __attribute__((__visibility__("default"))) void *
 ConflictSet_create(int64_t oldestVersion) {
-  mallocBytesDelta = 0;
+  return new (safe_malloc(sizeof(ConflictSet::Impl)))
-  auto *result = new (safe_malloc(sizeof(ConflictSet::Impl)))
       ConflictSet::Impl{oldestVersion};
-  result->totalBytes += mallocBytesDelta;
-  return result;
 }
 __attribute__((__visibility__("default"))) void ConflictSet_destroy(void *cs) {
   using Impl = ConflictSet::Impl;
   ((Impl *)cs)->~Impl();
-  safe_free(cs, sizeof(Impl));
+  free(cs);
-}
-__attribute__((__visibility__("default"))) int64_t
-ConflictSet_getBytes(void *cs) {
-  using Impl = ConflictSet::Impl;
-  return ((Impl *)cs)->totalBytes;
 }
 }
-
-#if SHOW_MEMORY
-struct __attribute__((visibility("default"))) PeakPrinter {
-  ~PeakPrinter() {
-    printf("--- skip_list ---\n");
-    printf("malloc bytes: %g\n", double(mallocBytes));
-    printf("Peak malloc bytes: %g\n", double(peakMallocBytes));
-  }
-} peakPrinter;
-#endif

TestDriver.cpp

@@ -5,6 +5,7 @@
 
 int main(int argc, char **argv) {
   for (int i = 1; i < argc; ++i) {
+    printf("Running: %s\n", argv[i]);
     std::ifstream t(argv[i], std::ios::binary);
     std::stringstream buffer;
     buffer << t.rdbuf();

aarch64-symbol-imports.txt

@@ -1,8 +0,0 @@
-__stack_chk_fail@GLIBC_2.17
-__stack_chk_guard@GLIBC_2.17
-abort@GLIBC_2.17
-free@GLIBC_2.17
-malloc@GLIBC_2.17
-memcpy@GLIBC_2.17
-memmove@GLIBC_2.17
-memset@GLIBC_2.17

apple-symbol-exports.txt

@@ -1,17 +0,0 @@
-_ConflictSet_addWrites
-_ConflictSet_check
-_ConflictSet_create
-_ConflictSet_destroy
-_ConflictSet_getBytes
-_ConflictSet_setOldestVersion
-__ZN8weaselab11ConflictSet16setOldestVersionEx
-__ZN8weaselab11ConflictSet9addWritesEPKNS0_10WriteRangeEix
-__ZN8weaselab11ConflictSetC1EOS0_
-__ZN8weaselab11ConflictSetC1Ex
-__ZN8weaselab11ConflictSetC2EOS0_
-__ZN8weaselab11ConflictSetC2Ex
-__ZN8weaselab11ConflictSetD1Ev
-__ZN8weaselab11ConflictSetD2Ev
-__ZN8weaselab11ConflictSetaSEOS0_
-__ZNK8weaselab11ConflictSet5checkEPKNS0_9ReadRangeEPNS0_6ResultEi
-__ZNK8weaselab11ConflictSet8getBytesEv

apple-symbol-imports.txt

@@ -1,8 +0,0 @@
-__tlv_bootstrap
-_abort
-_bzero
-_free
-_malloc
-_memcpy
-_memmove
-dyld_stub_binder

conflict_set.py

@@ -1,136 +0,0 @@
-import ctypes
-import enum
-import os
-
-from typing import Optional
-
-
-class _Key(ctypes.Structure):
-    _fields_ = [("p", ctypes.POINTER(ctypes.c_ubyte)), ("len", ctypes.c_int)]
-
-
-class ReadRange(ctypes.Structure):
-    _fields_ = [
-        ("begin", _Key),
-        ("end", _Key),
-        ("readVersion", ctypes.c_int64),
-    ]
-
-
-class WriteRange(ctypes.Structure):
-    _fields_ = [("begin", _Key), ("end", _Key)]
-
-
-class Result(enum.Enum):
-    COMMIT = 0
-    CONFLICT = 1
-    TOO_OLD = 2
-
-
-def write(begin: bytes, end: Optional[bytes] = None) -> WriteRange:
-    b = (ctypes.c_ubyte * len(begin)).from_buffer(bytearray(begin))
-
-    if end is None:
-        e = (ctypes.c_ubyte * 0)()
-    else:
-        e = (ctypes.c_ubyte * len(end)).from_buffer(bytearray(end))
-    return WriteRange(_Key(b, len(b)), _Key(e, len(e)))
-
-
-def read(version: int, begin: bytes, end: Optional[bytes] = None) -> ReadRange:
-    b = (ctypes.c_ubyte * len(begin)).from_buffer(bytearray(begin))
-    if end is None:
-        e = (ctypes.c_ubyte * 0)()
-    else:
-        e = (ctypes.c_ubyte * len(end)).from_buffer(bytearray(end))
-    return ReadRange(_Key(b, len(b)), _Key(e, len(e)), version)
-
-
-class ConflictSet:
-    def __init__(
-        self,
-        version: int = 0,
-        build_dir: Optional[str] = None,
-        implementation: Optional[str] = None,
-    ) -> None:
-        self._lib = None
-        if build_dir is None:
-            build_dir = os.path.dirname(__file__) + "/build"
-        if implementation is None:
-            implementation = "radix_tree"
-        for f in (
-            build_dir + "/" + implementation + "/libconflict-set.so.0",
-            os.path.dirname(__file__)
-            + "/build/"
-            + implementation
-            + "/libconflict-set.0.dylib",
-        ):
-            try:
-                self._lib = ctypes.cdll.LoadLibrary(f)
-            except:
-                pass
-
-        if self._lib is None:
-            import sys
-
-            print(
-                "Could not find libconflict-set implementation " + implementation,
-                file=sys.stderr,
-            )
-            sys.exit(1)
-
-        self._lib.ConflictSet_create.argtypes = (ctypes.c_int64,)
-        self._lib.ConflictSet_create.restype = ctypes.c_void_p
-
-        self._lib.ConflictSet_check.argtypes = (
-            ctypes.c_void_p,
-            ctypes.POINTER(ReadRange),
-            ctypes.POINTER(ctypes.c_int),
-            ctypes.c_int,
-        )
-
-        self._lib.ConflictSet_addWrites.argtypes = (
-            ctypes.c_void_p,
-            ctypes.POINTER(WriteRange),
-            ctypes.c_int,
-            ctypes.c_int64,
-        )
-
-        self._lib.ConflictSet_setOldestVersion.argtypes = (
-            ctypes.c_void_p,
-            ctypes.c_int64,
-        )
-
-        self._lib.ConflictSet_destroy.argtypes = (ctypes.c_void_p,)
-
-        self._lib.ConflictSet_getBytes.argtypes = (ctypes.c_void_p,)
-        self._lib.ConflictSet_getBytes.restype = ctypes.c_int64
-
-        self.p = self._lib.ConflictSet_create(version)
-
-    def addWrites(self, version: int, *writes: WriteRange):
-        self._lib.ConflictSet_addWrites(
-            self.p, (WriteRange * len(writes))(*writes), len(writes), version
-        )
-
-    def check(self, *reads: ReadRange) -> list[Result]:
-        r = (ctypes.c_int * len(reads))()
-        self._lib.ConflictSet_check(self.p, *reads, r, 1)
-        return [Result(x) for x in r]
-
-    def setOldestVersion(self, version: int) -> None:
-        self._lib.ConflictSet_setOldestVersion(self.p, version)
-
-    def getBytes(self) -> int:
-        return self._lib.ConflictSet_getBytes(self.p)
-
-    def __enter__(self):
-        return self
-
-    def close(self) -> None:
-        if self.p is not None:
-            self._lib.ConflictSet_destroy(self.p)
-            self.p = None
-
-    def __exit__(self, exception_type, exception_value, exception_traceback):
-        self.close()

conflict_set_c_api_test.c

@@ -7,7 +7,6 @@ int main(void) {
   ConflictSet_WriteRange w;
   ConflictSet_Result result;
   ConflictSet_ReadRange r;
-  int64_t bytes;
   w.begin.p = (const uint8_t *)"0000";
   w.begin.len = 4;
   w.end.len = 0;
@@ -18,8 +17,6 @@ int main(void) {
   r.readVersion = 0;
   ConflictSet_check(cs, &r, &result, 1);
   assert(result == ConflictSet_Conflict);
-  bytes = ConflictSet_getBytes(cs);
-  assert(bytes > 0);
   ConflictSet_destroy(cs);
   return 0;
 }

conflict_set_cxx_api_test.cpp

@@ -2,8 +2,6 @@
 
 #include <cassert>
 
-using namespace weaselab;
-
 int main(void) {
   ConflictSet cs(0);
   ConflictSet::WriteRange w;
@@ -19,6 +17,4 @@ int main(void) {
   r.readVersion = 0;
   cs.check(&r, &result, 1);
   assert(result == ConflictSet::Conflict);
-  int64_t bytes = cs.getBytes();
-  assert(bytes > 0);
 }