Work around a compiler bug in MacOS AppleClang 11 (#6103)

* Workaround a compiler bug in MacOS AppleClang

* [CI] Run C++ test with MacOS Catalina + AppleClang 11.0.3

* [CI] Migrate cmake_test on MacOS from Travis CI to GitHub Actions

* Install OpenMP runtime

* [CI] Use CMake to locate lz4 lib

.github/workflows/main.yml

@@ -11,6 +11,30 @@ env:
 
 # A workflow run is made up of one or more jobs that can run sequentially or in parallel
 jobs:
+  gtest-cpu:
+    name: Test Google C++ test (CPU)
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [macos-10.15]
+    steps:
+    - uses: actions/checkout@v2
+      with:
+        submodules: 'true'
+    - name: Install system packages
+      run: |
+        brew install lz4 ninja libomp
+    - name: Build gtest binary
+      run: |
+        mkdir build
+        cd build
+        cmake .. -DGOOGLE_TEST=ON -DUSE_OPENMP=ON -DUSE_DMLC_GTEST=ON -DPLUGIN_LZ4=ON -DPLUGIN_DENSE_PARSER=ON -GNinja
+        ninja -v
+    - name: Run gtest binary
+      run: |
+        cd build
+        ctest --extra-verbose
   test-with-jvm:
     name: Test JVM on OS ${{ matrix.os }}
     runs-on: ${{ matrix.os }}

.travis.yml

@@ -1,6 +1,5 @@
 sudo: required
 
-osx_image: xcode10.1
 dist: bionic
 
 env:
@@ -15,16 +14,16 @@ jobs:
       env: TASK=python_sdist_test
     - os: osx
       arch: amd64
+      osx_image: xcode10.1
       env: TASK=python_test
     - os: osx
       arch: amd64
+      osx_image: xcode10.1
       env: TASK=python_sdist_test
     - os: osx
       arch: amd64
+      osx_image: xcode10.1
       env: TASK=java_test
-    - os: osx
-      arch: amd64
-      env: TASK=cmake_test
     - os: linux
       arch: s390x
       env: TASK=s390x_test

plugin/CMakeLists.txt

@@ -1,6 +1,11 @@
 if (PLUGIN_LZ4)
   target_sources(objxgboost PRIVATE ${xgboost_SOURCE_DIR}/plugin/lz4/sparse_page_lz4_format.cc)
-  target_link_libraries(objxgboost PUBLIC lz4)
+  find_path(LIBLZ4_INCLUDE_DIR lz4.h)
+  find_library(LIBLZ4_LIBRARY NAMES lz4)
+  message(STATUS "LIBLZ4_INCLUDE_DIR = ${LIBLZ4_INCLUDE_DIR}")
+  message(STATUS "LIBLZ4_LIBRARY = ${LIBLZ4_LIBRARY}")
+  target_include_directories(objxgboost PUBLIC ${LIBLZ4_INCLUDE_DIR})
+  target_link_libraries(objxgboost PUBLIC ${LIBLZ4_LIBRARY})
 endif (PLUGIN_LZ4)
 
 if (PLUGIN_DENSE_PARSER)

src/common/threading_utils.h

@@ -109,7 +109,7 @@ class BlockedSpace2d {
 
 
 // Wrapper to implement nested parallelism with simple omp parallel for
-template<typename Func>
+template <typename Func>
 void ParallelFor2d(const BlockedSpace2d& space, int nthreads, Func func) {
   const size_t num_blocks_in_space = space.Size();
   nthreads = std::min(nthreads, omp_get_max_threads());
@@ -118,7 +118,8 @@ void ParallelFor2d(const BlockedSpace2d& space, int nthreads, Func func) {
   dmlc::OMPException omp_exc;
 #pragma omp parallel num_threads(nthreads)
   {
-    omp_exc.Run([&]() {
+    omp_exc.Run(
+        [](size_t num_blocks_in_space, const BlockedSpace2d& space, int nthreads, Func func) {
       size_t tid = omp_get_thread_num();
       size_t chunck_size =
           num_blocks_in_space / nthreads + !!(num_blocks_in_space % nthreads);
@@ -128,7 +129,7 @@ void ParallelFor2d(const BlockedSpace2d& space, int nthreads, Func func) {
       for (auto i = begin; i < end; i++) {
         func(space.GetFirstDimension(i), space.GetRange(i));
       }
-    });
+    }, num_blocks_in_space, space, nthreads, func);
   }
   omp_exc.Rethrow();
 }

tests/travis/run_test.sh

@@ -17,6 +17,12 @@ if [ ${TASK} == "python_sdist_test" ]; then
 fi
 
 if [ ${TASK} == "python_test" ]; then
+    if grep -n -R '<<<.*>>>\(.*\)' src include | grep --invert "NOLINT"; then
+        echo 'Do not use raw CUDA execution configuration syntax with <<<blocks, threads>>>.' \
+             'try `dh::LaunchKernel`'
+        exit -1
+    fi
+
     set -e
     # Build/test
     rm -rf build
@@ -69,26 +75,6 @@ if [ ${TASK} == "java_test" ]; then
     mvn -q test
 fi
 
-if [ ${TASK} == "cmake_test" ]; then
-    set -e
-
-    if grep -n -R '<<<.*>>>\(.*\)' src include | grep --invert "NOLINT"; then
-        echo 'Do not use raw CUDA execution configuration syntax with <<<blocks, threads>>>.' \
-             'try `dh::LaunchKernel`'
-        exit -1
-    fi
-
-    # Build/test
-    rm -rf build
-    mkdir build && cd build
-    PLUGINS="-DPLUGIN_LZ4=ON -DPLUGIN_DENSE_PARSER=ON"
-    cmake .. -DCMAKE_VERBOSE_MAKEFILE=ON -DGOOGLE_TEST=ON -DUSE_OPENMP=ON -DUSE_DMLC_GTEST=ON ${PLUGINS}
-    make -j$(nproc)
-    ./testxgboost
-    cd ..
-    rm -rf build
-fi
-
 if [ ${TASK} == "s390x_test" ]; then
     set -e
 

tests/travis/setup.sh

@@ -16,10 +16,6 @@ if [ ${TASK} == "python_test" ] || [ ${TASK} == "python_sdist_test" ]; then
     conda create -n python3 python=3.7
 fi
 
-if [ ${TASK} == "cmake_test" ] && [ ${TRAVIS_OS_NAME} == "osx" ]; then
-    sudo softwareupdate -i "Command Line Tools (macOS High Sierra version 10.13) for Xcode-9.3"
-fi
-
 if [ ${TASK} == "s390x_test" ] && [ ${TRAVIS_CPU_ARCH} == "s390x" ]; then
     sudo snap install cmake --channel=3.17/beta --classic
     export PATH=/snap/bin:${PATH}