Bring in a bunch of cmake settings from conflict-set

2025-05-23 12:43:25 -04:00
parent 1217ded8a7
commit 4240894318
4 changed files with 79 additions and 13 deletions
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -7,6 +7,11 @@ project(
  LANGUAGES C CXX)
 set(CMAKE_CXX_STANDARD 20)

+include(CMakePushCheckState)
+include(CheckCXXCompilerFlag)
+include(CheckIncludeFileCXX)
+include(CheckCXXSourceCompiles)
+
 set(DEFAULT_BUILD_TYPE "Release")

 if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
@@ -28,7 +33,55 @@ add_compile_options(
  -fPIC
  -fdata-sections
  -ffunction-sections
-  -fno-omit-frame-pointer)
+  -fno-omit-frame-pointer
+  -g)
+
+if(CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
+  add_link_options("-Wno-unused-command-line-argument")
+  find_program(LLVM_OBJCOPY llvm-objcopy)
+  if(LLVM_OBJCOPY)
+    set(CMAKE_OBJCOPY
+        ${LLVM_OBJCOPY}
+        CACHE FILEPATH "path to objcopy binary" FORCE)
+  endif()
+endif()
+
+if(CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+  add_compile_options("-Wno-maybe-uninitialized")
+endif()
+
+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()
+
+if(CMAKE_SYSTEM_PROCESSOR STREQUAL aarch64 OR CMAKE_SYSTEM_PROCESSOR STREQUAL
+                                              arm64)
+  add_compile_options(-mbranch-protection=standard)
+else()
+  add_compile_options(-fcf-protection)
+  set(rewrite_endbr_flags "-fuse-ld=mold;LINKER:-z,rewrite-endbr")
+  cmake_push_check_state()
+  list(APPEND CMAKE_REQUIRED_LINK_OPTIONS ${rewrite_endbr_flags})
+  check_cxx_source_compiles("int main(){}" HAS_REWRITE_ENDBR FAIL_REGEX
+                            "warning:")
+  if(HAS_REWRITE_ENDBR)
+    add_link_options(${rewrite_endbr_flags})
+  endif()
+  cmake_pop_check_state()
+endif()
+
+set(version_script_flags
+    LINKER:--version-script=${CMAKE_CURRENT_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()

 if(CMAKE_SYSTEM_PROCESSOR STREQUAL x86_64)
  add_compile_options(-mavx)
@@ -37,6 +90,16 @@ if(CMAKE_SYSTEM_NAME STREQUAL Linux)
  add_link_options(LINKER:--gc-sections)
 endif()

+# 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)
+
+if(HAS_VERSION_SCRIPT)
+  target_link_options(
+    ${PROJECT_NAME} PRIVATE
+    LINKER:--version-script=${CMAKE_CURRENT_SOURCE_DIR}/linker.map)
+endif()
+
 add_subdirectory(third_party)

 set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_SOURCE_DIR}/cmake/")
@@ -93,9 +156,12 @@ if(NOT APPLE)
      "Proceeding with all symbols global in static library")
 endif()

+set(TEST_FLAGS -Wall -Wextra -Wunreachable-code -Wpedantic -UNDEBUG)
+
 add_executable(mytest src/test.cpp)
 target_include_directories(mytest PRIVATE include)
 target_link_libraries(mytest PRIVATE ${PROJECT_NAME} doctest nanobench simdjson)
+target_compile_options(mytest PRIVATE ${TEST_FLAGS})
 doctest_discover_tests(mytest)

 include(CMakePushCheckState)
@@ -109,7 +175,7 @@ if(HAS_LIB_FUZZER)
  add_executable(fuzz src/fuzz.cpp src/lib.cpp)
  target_include_directories(fuzz PRIVATE include)
  target_link_libraries(fuzz PRIVATE simdjson)
-  target_compile_options(fuzz PRIVATE -fsanitize=fuzzer)
+  target_compile_options(fuzz PRIVATE -fsanitize=fuzzer ${TEST_FLAGS})
  target_link_options(fuzz PRIVATE -fsanitize=fuzzer)
 endif()

--- a/src/callbacks.h
+++ b/src/callbacks.h
@@ -34,12 +34,12 @@ inline WeaselJsonCallbacks noopCallbacks() {
  result.on_begin_object = +[](void *) {};
  result.on_end_object = +[](void *) {};
  result.on_begin_string = +[](void *) {};
-  result.on_string_data = +[](void *, const char *buf, int len) {};
+  result.on_string_data = +[](void *, const char *, int) {};
  result.on_end_string = +[](void *) {};
  result.on_begin_array = +[](void *) {};
  result.on_end_array = +[](void *) {};
  result.on_begin_number = +[](void *) {};
-  result.on_number_data = +[](void *, const char *buf, int len) {};
+  result.on_number_data = +[](void *, const char *, int) {};
  result.on_end_number = +[](void *) {};
  result.on_true_literal = +[](void *) {};
  result.on_false_literal = +[](void *) {};
--- a/src/parser3.h
+++ b/src/parser3.h
@@ -550,7 +550,7 @@ inline PRESERVE_NONE WeaselJsonStatus n_string_following_escape(Parser3 *self) {
  case 'n':
  case 'r':
  case 't':
-    *self->writeBuf++ = tables.unescape[*self->buf++];
+    *self->writeBuf++ = tables.unescape[uint8_t(*self->buf++)];
    self->pop();
    if (auto s = self->push({N_STRING2})) {
      return s;
--- a/src/test.cpp
+++ b/src/test.cpp
@@ -153,7 +153,7 @@ void testStreaming(std::string const &json) {
  {
    auto copy = json;
    auto *parser = WeaselJsonParser_create(1024, &c, &streaming);
-    for (int i = 0; i < copy.size(); ++i) {
+    for (size_t i = 0; i < copy.size(); ++i) {
      REQUIRE(WeaselJsonParser_parse(parser, copy.data() + i, 1) ==
              WeaselJson_AGAIN);
    }
@@ -179,7 +179,7 @@ TEST_CASE("parser3") {
  {
    auto copy = json;
    auto *parser = WeaselJsonParser_create(1024, &c, &state);
-    for (int i = 0; i < copy.size(); ++i) {
+    for (size_t i = 0; i < copy.size(); ++i) {
      REQUIRE(WeaselJsonParser_parse(parser, copy.data() + i, 1) ==
              WeaselJson_AGAIN);
    }
@@ -189,7 +189,7 @@ TEST_CASE("parser3") {
  {
    std::string copy = "{\"x\": [], \"y\": {}}";
    auto *parser = WeaselJsonParser_create(1024, &c, &state);
-    for (int i = 0; i < copy.size(); ++i) {
+    for (size_t i = 0; i < copy.size(); ++i) {
      REQUIRE(WeaselJsonParser_parse(parser, copy.data() + i, 1) ==
              WeaselJson_AGAIN);
    }
@@ -201,7 +201,7 @@ TEST_CASE("parser3") {
    auto c = noopCallbacks();
    std::string copy = "{\"a\":\"a";
    auto *parser = WeaselJsonParser_create(1024, &c, &state);
-    for (int i = 0; i < copy.size(); ++i) {
+    for (size_t i = 0; i < copy.size(); ++i) {
      REQUIRE(WeaselJsonParser_parse(parser, copy.data() + i, 1) ==
              WeaselJson_AGAIN);
    }
@@ -212,7 +212,7 @@ TEST_CASE("parser3") {
    auto c = noopCallbacks();
    std::string copy = "[";
    auto *parser = WeaselJsonParser_create(1024, &c, &state);
-    for (int i = 0; i < copy.size(); ++i) {
+    for (size_t i = 0; i < copy.size(); ++i) {
      REQUIRE(WeaselJsonParser_parse(parser, copy.data() + i, 1) ==
              WeaselJson_AGAIN);
    }
@@ -236,7 +236,7 @@ void doTestUnescapingUtf8(std::string const &escaped,
  };
  auto *parser = WeaselJsonParser_create(1024, &c, &result);
  auto copy = escaped;
-  for (int i = 0; i < copy.size(); i += stride) {
+  for (size_t i = 0; i < copy.size(); i += stride) {
    CAPTURE(i);
    REQUIRE(WeaselJsonParser_parse(parser, copy.data() + i,
                                   std::min<int>(stride, copy.size() - i)) ==
@@ -282,11 +282,11 @@ TEST_CASE("bench3") {
  bench.batch(json.size());
  bench.unit("byte");
  auto *parser = WeaselJsonParser_create(1024, &c, nullptr);
-  for (int stride = 1; stride <= json.size(); stride *= 2) {
+  for (size_t stride = 1; stride <= json.size(); stride *= 2) {
    bench.run("parser3 (stride: " + std::to_string(stride) + ")", [&]() {
      auto copy = json;
      WeaselJsonParser_reset(parser);
-      for (int i = 0; i < copy.size(); i += stride) {
+      for (size_t i = 0; i < copy.size(); i += stride) {
        if (WeaselJsonParser_parse(parser, copy.data() + i,
                                   std::min<int>(copy.size() - i, stride)) !=
            WeaselJson_AGAIN) {