Draft-Release.yml

# Perform thorough builds, and potentially create a draft CI workflow. 
name: Draft Release

on:
  # Trigger on pushes to matching tags.
  push:
    tags:
      - 'v*.*.*'
  # Or on pull_requests which mutate this CI workflow, using pr rather than push:branch to ensure that the merged state will be OK, as that is what is important here.
  pull_request:
    paths:
      - ".github/workflows/Draft-Release.yml"
  # Or trigger on manual dispatch. This will not produce a release, but will perform the thorough build. 
  workflow_dispatch:

defaults:
  run:
    # Default to using bash regardless of OS unless otherwise specified.
    shell: bash

# @todo - add a job/step which validates the version is of the correct format (for pre releases) and that the git tag matches the CMake/header version(s) 

# Several jobs with some dependencies between them.
# + Thorough Ubuntu builds
#   + Oldest and newest cuda, lots of arch, vis off, tests on
# + Thorough Windows builds
#   + Oldest and newest cuda, lots of arch, vis off, tests on
# + Wheel producing manylinux builds
#   + CUDA 11.2 and 12.0, py 3.8-3.12, vis on/off, py only.
# + Wheel producing Windows builds
#   + CUDA 11.2 and 12.0, py 3.8-3.12, vis on/off, py only.
# + Draft github release workflow.

jobs:
  build-ubuntu:
    runs-on: ${{ matrix.cudacxx.os }}
    strategy:
      fail-fast: false
      # Multiplicative build matrix
      # optional exclude: can be partial, include: must be specific
      matrix:
        # CUDA_ARCH values are reduced compared to wheels due to CI memory issues while compiling the test suite.
        cudacxx:
          - cuda: "12.6"
            cuda_arch: "50-real;90-real;90-virtual;"
            hostcxx: gcc-12
            os: ubuntu-22.04
          - cuda: "12.0"
            cuda_arch: "50-real;90-real;90-virtual;"
            hostcxx: gcc-11
            os: ubuntu-22.04
          - cuda: "11.8"
            cuda_arch: "35-real;90-real;90-virtual"
            hostcxx: gcc-9
            os: ubuntu-20.04
          - cuda: "11.0"
            cuda_arch: "35-real;80-real;80-virtual"
            hostcxx: gcc-8
            os: ubuntu-20.04
        python:
          - "3.8"
        config:
          - name: "Release"
            config: "Release"
            SEATBELTS: "ON"
        VISUALISATION: 
          - "OFF"

    # Name the job based on matrix/env options
    name: "build-ubuntu (${{ matrix.cudacxx.cuda }}, ${{matrix.python}}, ${{ matrix.VISUALISATION }}, ${{ matrix.config.name }}, ${{ matrix.cudacxx.os }})"

    # Define job-wide env constants, and promote matrix elements to env constants for portable steps.
    env:
      # Define constants
      BUILD_DIR: "build"
      FLAMEGPU_BUILD_TESTS: "ON"
      # Conditional based on matrix via awkward almost ternary
      FLAMEGPU_BUILD_PYTHON: ${{ fromJSON('{true:"ON",false:"OFF"}')[matrix.python != ''] }}
      # Port matrix options to environment, for more portability.
      CUDA: ${{ matrix.cudacxx.cuda }}
      CUDA_ARCH: ${{ matrix.cudacxx.cuda_arch }}
      HOSTCXX: ${{ matrix.cudacxx.hostcxx }}
      OS: ${{ matrix.cudacxx.os }}
      CONFIG: ${{ matrix.config.config }}
      FLAMEGPU_SEATBELTS: ${{ matrix.config.SEATBELTS }}
      PYTHON: ${{ matrix.python}}
      VISUALISATION: ${{ matrix.VISUALISATION }}

    steps:
    - uses: actions/checkout@v3

    - name: Install CUDA
      if: ${{ startswith(env.OS, 'ubuntu') && env.CUDA != '' }}
      env:
        cuda: ${{ env.CUDA }}
      run: .github/scripts/install_cuda_ubuntu.sh

    - name: Install/Select gcc and g++
      if: ${{ startsWith(env.HOSTCXX, 'gcc-') }}
      run: |
        gcc_version=${HOSTCXX//gcc-/}
        sudo apt-get install -y gcc-${gcc_version} g++-${gcc_version}
        echo "CC=/usr/bin/gcc-${gcc_version}" >> $GITHUB_ENV
        echo "CXX=/usr/bin/g++-${gcc_version}" >> $GITHUB_ENV
        echo "CUDAHOSTCXX=/usr/bin/g++-${gcc_version}" >> $GITHUB_ENV

    - name: Select Python
      if: ${{ env.PYTHON != '' && env.FLAMEGPU_BUILD_PYTHON == 'ON' }} 
      uses: actions/setup-python@v4
      with:
        python-version: ${{ env.PYTHON }}

    - name: Install python dependencies
      if: ${{ env.PYTHON != '' && env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      run: |
        sudo apt-get install python3-venv
        python3 -m pip install --upgrade wheel build setuptools

    - name: Install Visualisation Dependencies
      if: ${{ startswith(env.OS, 'ubuntu') && env.VISUALISATION == 'ON' }}
      run: |
        # Install ubuntu-22.04 packages
        if [ "$OS" == 'ubuntu-22.04' ]; then
          sudo apt-get install -y libglew-dev libfontconfig1-dev libsdl2-dev libdevil-dev libfreetype-dev
        fi
        # Install ubuntu-20.04 packages
        if [ "$OS" == 'ubuntu-20.04' ]; then
          sudo apt-get install -y libglew-dev libfontconfig1-dev libsdl2-dev libdevil-dev libfreetype-dev
        fi
        # Install Ubuntu 18.04 packages
        if [ "$OS" == 'ubuntu-18.04' ]; then
          sudo apt-get install -y libglew-dev libfontconfig1-dev libsdl2-dev libdevil-dev libfreetype6-dev libgl1-mesa-dev
        fi
    
    - name: Install Swig >= 4.0.2
      run: |
        # Remove existing swig install, so CMake finds the correct swig
        if [ "$OS" == 'ubuntu-20.04' ]; then
          sudo apt-get remove -y swig swig4.0
        fi
        if [ "$OS" == 'ubuntu-18.04' ]; then
          sudo apt-get remove -y swig
        fi
        # Install additional apt-based dependencies required to build swig 4.0.2
        sudo apt-get install -y bison
        # Create a local directory to build swig in.
        mkdir -p swig-from-source && cd swig-from-source
        # Install SWIG building from source dependencies
        wget https://github.com/swig/swig/archive/refs/tags/v4.0.2.tar.gz
        tar -zxf v4.0.2.tar.gz
        cd swig-4.0.2/
        ./autogen.sh
        ./configure
        make
        sudo make install

    # This pre-emptively patches a bug where ManyLinux didn't generate buildnumber as git dir was owned by diff user
    - name: Enable git safe-directory
      run: git config --global --add safe.directory $GITHUB_WORKSPACE

    - name: Configure cmake
      run: >
        cmake . -B "${{ env.BUILD_DIR }}"
        -DCMAKE_BUILD_TYPE="${{ env.CONFIG }}"
        -Werror=dev
        -DCMAKE_WARN_DEPRECATED="OFF" 
        -DFLAMEGPU_WARNINGS_AS_ERRORS="ON"
        -DCMAKE_CUDA_ARCHITECTURES="${{ env.CUDA_ARCH }}"
        -DFLAMEGPU_BUILD_TESTS="${{ env.FLAMEGPU_BUILD_TESTS }}"
        -DFLAMEGPU_BUILD_PYTHON="${{ env.FLAMEGPU_BUILD_PYTHON }}"
        -DPYTHON3_EXACT_VERSION="${{ env.PYTHON }}"
        -DFLAMEGPU_VISUALISATION="${{ env.VISUALISATION }}"
        -DFLAMEGPU_ENABLE_NVTX="ON"

    - name: Build static library
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --target flamegpu --verbose -j `nproc`

    - name: Build python wheel
      if: ${{ env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --target pyflamegpu --verbose -j `nproc`

    - name: Build tests
      if: ${{ env.FLAMEGPU_BUILD_TESTS == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --target tests --verbose -j `nproc`

    - name: Build all remaining targets
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --target all --verbose -j `nproc`


  # Thorough Windows builds
  build-windows:
    runs-on: ${{ matrix.cudacxx.os }}
    strategy:
      fail-fast: false
      # Multiplicative build matrix
      # optional exclude: can be partial, include: must be specific
      matrix:
        # CUDA_ARCH values are reduced compared to wheels due to CI memory issues while compiling the test suite.
        cudacxx:
          - cuda: "12.6.0"
            cuda_arch: "50-real;90-real;90-virtual"
            hostcxx: "Visual Studio 17 2022"
            os: windows-2022
          - cuda: "12.0.0"
            cuda_arch: "50-real;90-real;90-virtual"
            hostcxx: "Visual Studio 16 2019"
            os: windows-2019
          - cuda: "11.8.0"
            cuda_arch: "35-real;90-real;90-virtual"
            hostcxx: "Visual Studio 16 2019"
            os: windows-2019
          - cuda: "11.0.3"
            cuda_arch: "35-real;80-real;80-virtual"
            hostcxx: "Visual Studio 16 2019"
            os: windows-2019
        python: 
          - "3.8"
        config:
          - name: "Release"
            config: "Release"
            SEATBELTS: "ON"
        VISUALISATION: 
          - "OFF"

    # Name the job based on matrix/env options
    name: "build-windows (${{ matrix.cudacxx.cuda }}, ${{matrix.python}}, ${{ matrix.VISUALISATION }}, ${{ matrix.config.name }}, ${{ matrix.cudacxx.os }})"

    # Define job-wide env constants, and promote matrix elements to env constants for portable steps.
    env:
      # Define constants
      BUILD_DIR: "build"
      FLAMEGPU_BUILD_TESTS: "ON"
      # Conditional based on matrix via awkward almost ternary
      FLAMEGPU_BUILD_PYTHON: ${{ fromJSON('{true:"ON",false:"OFF"}')[matrix.python != ''] }}
      # Port matrix options to environment, for more portability.
      CUDA: ${{ matrix.cudacxx.cuda }}
      CUDA_ARCH: ${{ matrix.cudacxx.cuda_arch }}
      HOSTCXX: ${{ matrix.cudacxx.hostcxx }}
      OS: ${{ matrix.cudacxx.os }}
      CONFIG: ${{ matrix.config.config }}
      FLAMEGPU_SEATBELTS: ${{ matrix.config.SEATBELTS }}
      PYTHON: ${{ matrix.python}}
      VISUALISATION: ${{ matrix.VISUALISATION }}
      # Ensure MSVC >= 1940 works with CUDA <= 12.3
      CUDAFLAGS: -allow-unsupported-compiler

    steps:
    - uses: actions/checkout@v3

    - name: Install CUDA (Windows)
      if: ${{ runner.os == 'Windows' && env.CUDA != '' }}
      shell: powershell
      env:
        cuda: ${{ env.CUDA }}
        visual_studio: ${{ env.HOSTCXX }}
      run: .github\scripts\install_cuda_windows.ps1

    - name: Select Python
      if: ${{ env.PYTHON != '' && env.FLAMEGPU_BUILD_PYTHON == 'ON' }} 
      uses: actions/setup-python@v4
      with:
        python-version: ${{ env.PYTHON }}
    
    - name: Install python dependencies
      if: ${{ env.PYTHON != '' && env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      run: |
        python3 -m pip install --upgrade wheel build setuptools

    - name: Add custom problem matchers for annotations
      run: echo "::add-matcher::.github/problem-matchers.json"

    # This pre-emptively patches a bug where ManyLinux didn't generate buildnumber as git dir was owned by diff user
    - name: Enable git safe-directory
      run: git config --global --add safe.directory $GITHUB_WORKSPACE
      
    # Must pass -G -A for windows, and -DPython3_ROOT_DIR/-DPYTHON3_EXECUTABLE as a github action workaround
    - name: Configure cmake
      run: >
        cmake . -B "${{ env.BUILD_DIR }}" 
        -G "${{ env.HOSTCXX }}" -A x64
        -Werror=dev
        -DCMAKE_WARN_DEPRECATED="OFF"
        -DFLAMEGPU_WARNINGS_AS_ERRORS="ON"
        -DCMAKE_CUDA_ARCHITECTURES="${{ env.CUDA_ARCH }}"
        -DFLAMEGPU_BUILD_TESTS="${{ env.FLAMEGPU_BUILD_TESTS }}"
        -DFLAMEGPU_BUILD_PYTHON="${{ env.FLAMEGPU_BUILD_PYTHON }}" 
        -DPython3_ROOT_DIR="$(dirname $(which python))" 
        -DPython3_EXECUTABLE="$(which python)"
        -DFLAMEGPU_VISUALISATION="${{ env.VISUALISATION }}"
        -DFLAMEGPU_ENABLE_NVTX="ON"

    - name: Build static library
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --config ${{ env.CONFIG }} --target flamegpu --verbose -j `nproc`

    - name: Build python wheel
      if: ${{ env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --config ${{ env.CONFIG }} --target pyflamegpu --verbose -j `nproc`

    - name: Build tests
      if: ${{ env.FLAMEGPU_BUILD_TESTS == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --config ${{ env.CONFIG }} --target tests --verbose -j `nproc`

    - name: Build all remaining targets
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --config ${{ env.CONFIG }} --target ALL_BUILD --verbose -j `nproc`


  # Manylinux2014 Wheel builds, using the manylinux2014 container
  wheel-manylinux2014:
    runs-on: ${{ matrix.cudacxx.os }}
    # Run steps inside a manylinux container.
    container: quay.io/pypa/manylinux2014_x86_64
    strategy:
      fail-fast: false
      # Multiplicative build matrix
      # optional exclude: can be partial, include: must be specific
      matrix:
        cudacxx:
          - cuda: "12.0"
            cuda_arch: "50-real;60-real;70-real;80-real;90-real;90-virtual"
            hostcxx: devtoolset-10
            os: ubuntu-20.04
          - cuda: "11.2"
            cuda_arch: "35-real;50-real;60-real;70-real;80-real;80-virtual"
            hostcxx: devtoolset-9
            os: ubuntu-20.04
        python: 
          - "3.12"
          - "3.11"
          - "3.10"
          - "3.9"
          - "3.8"
        config:
          - name: "Release"
            config: "Release"
            SEATBELTS: "ON"
        VISUALISATION: 
          - "ON"
          - "OFF"

    # Name the job based on matrix/env options
    name: "wheel-manylinux2014 (${{ matrix.cudacxx.cuda }}, ${{matrix.python}}, ${{ matrix.VISUALISATION }}, ${{ matrix.config.name }}, ${{ matrix.cudacxx.os }})"

    env:
      # Control if the wheel should be repaired. This will fail until .so's are addressed
      AUDITWHEEL_REPAIR: "OFF"
      MANYLINUX: "manylinux2014"
      ARCH: "x86_64"
      # Control if static GLEW should be built and used or not.
      USE_STATIC_GLEW: "ON"
      # Compute the wheelhouse name which should be unique within the matrix. This must be unique per build matrix/job combination
      ARTIFACT_NAME: wheel-manylinux2014-${{ matrix.cudacxx.cuda }}-${{matrix.python}}-${{ matrix.VISUALISATION }}-${{ matrix.config.name }}-${{ matrix.cudacxx.os }}
      # Define constants
      BUILD_DIR: "build"
      FLAMEGPU_BUILD_TESTS: "OFF"
      # Conditional based on matrix via awkward almost ternary
      FLAMEGPU_BUILD_PYTHON: ${{ fromJSON('{true:"ON",false:"OFF"}')[matrix.python != ''] }}
      # Port matrix options to environment, for more portability.
      CUDA: ${{ matrix.cudacxx.cuda }}
      CUDA_ARCH: ${{ matrix.cudacxx.cuda_arch }}
      HOSTCXX: ${{ matrix.cudacxx.hostcxx }}
      OS: ${{ matrix.cudacxx.os }}
      CONFIG: ${{ matrix.config.config }}
      FLAMEGPU_SEATBELTS: ${{ matrix.config.SEATBELTS }}
      PYTHON: ${{ matrix.python}}
      VISUALISATION: ${{ matrix.VISUALISATION }}
      # Short term fix to use node16 not node20 for actions. This will stop working eventually, forcing our hand in dropping manylinux2014 support.
      ACTIONS_ALLOW_USE_UNSECURE_NODE_VERSION: true

    steps:
    - uses: actions/checkout@v3

    # Downgrade the devtoolset in the image based on the build matrix, using:
    # gcc-10 for CUDA >= 11.2. Unclear if devtoolset-10 will upgrade to unpatched 11.3 which breaks CUDA builds that use <chrono>. 
    # gcc-9 for CUDA >= 11.0
    # these are not the officially supported toolset on centos by cuda, but it's what works.
    - name: Install RHEL devtoolset (CentOS)
      if: ${{ startsWith(env.HOSTCXX, 'devtoolset-') }}
      run: |
        # Install devtoolset-X
        yum install -y ${{ env.HOSTCXX }}
        # Enable the toolset via source not scl enable which doesn't get on with multi-step GHA 
        source /opt/rh/${{ env.HOSTCXX }}/enable
        # Export the new environment / compilers for subsequent steps.
        echo "PATH=${PATH}" >> $GITHUB_ENV
        echo "CC=$(which gcc)" >> $GITHUB_ENV
        echo "CXX=$(which g++)" >> $GITHUB_ENV
        echo "CUDAHOSTCXX=$(which g++)" >> $GITHUB_ENV

    - name: Install CUDA (CentOS)
      if: ${{ env.CUDA != '' }}
      env:
        cuda: ${{ env.CUDA }}
      run: .github/scripts/install_cuda_centos.sh

    - name: Install Visualisation Dependencies (CentOS)
      if: ${{ env.VISUALISATION == 'ON' }}
      run: |
        yum install -y glew-devel fontconfig-devel SDL2-devel freetype-devel 
        # Build/Install DevIL from source.
        yum install -y freeglut-devel
        git clone --depth 1 https://github.com/DentonW/DevIL.git
        cd DevIL/DevIL
        mkdir -p build
        cd build
        cmake .. -DCMAKE_BUILD_TYPE=${{ env.CONFIG }} -Wno-error=dev
        make -j `nproc`
        make install

    - name: Build and install GLEW including static GLEW
      if: ${{ env.VISUALISATION == 'ON' && env.USE_STATIC_GLEW == 'ON' }}
      env:
        GLEW_VERSION: "2.2.0"
      run: |
        yum install -y wget
        wget https://github.com/nigels-com/glew/releases/download/glew-${{ env.GLEW_VERSION }}/glew-${{ env.GLEW_VERSION }}.tgz
        tar -zxf glew-${{ env.GLEW_VERSION }}.tgz
        cd glew-${{ env.GLEW_VERSION }}
        make
        make install

    # This patches a bug where ManyLinux doesn't generate buildnumber as git dir is owned by diff user
    - name: Enable git safe-directory
      run: git config --global --add safe.directory $GITHUB_WORKSPACE

    # Unlike other builds manylinux, uses static glew as it has been built and installed.
    - name: Configure cmake
      run: >
        cmake . -B "${{ env.BUILD_DIR }}"
        -DCMAKE_BUILD_TYPE="${{ env.CONFIG }}"
        -Werror=dev
        -DCMAKE_WARN_DEPRECATED="OFF" 
        -DFLAMEGPU_WARNINGS_AS_ERRORS="ON"
        -DCMAKE_CUDA_ARCHITECTURES="${{ env.CUDA_ARCH }}"
        -DFLAMEGPU_BUILD_TESTS="${{ env.FLAMEGPU_BUILD_TESTS }}"
        -DFLAMEGPU_BUILD_PYTHON="${{ env.FLAMEGPU_BUILD_PYTHON }}"
        -DPYTHON3_EXACT_VERSION="${{ env.PYTHON }}"
        -DFLAMEGPU_VISUALISATION="${{ env.VISUALISATION }}"
        -DFLAMEGPU_ENABLE_NVTX="ON"
        -DGLEW_USE_STATIC_LIBS="${{ env.USE_STATIC_GLEW }}"
        -DOpenGL_GL_PREFERENCE:STRING=LEGACY
        -DFLAMEGPU_BUILD_PYTHON_PATCHELF=ON

    - name: Build python wheel
      if: ${{ env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --target pyflamegpu --verbose -j `nproc`

    # Run audithweel show for information, but do not repair.
    - name: Run auditwheel show
      working-directory: ${{ env.BUILD_DIR }}
      run: auditwheel show lib/${{ env.CONFIG }}/python/dist/*whl

    # Ideally we should use auditwheel repair to check/enforce conformity
    # But we cannot due to cuda shared object (libcuda.so.1) dependencies which we cannot/shouldnot/wil not package into the wheel. 
    - name: Run auditwheel repair
      if: ${{ env.AUDITWHEEL_REPAIR == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: auditwheel repair --plat ${{ env.MANYLINUX }}_${{ env.ARCH }} lib/${{ env.CONFIG }}/python/dist/*whl -w lib/${{ env.CONFIG }}/python/dist

    # Upload wheel artifacts to the job on GHA, with a short retention
    # Use a unique name per job matrix run, to avoid a risk of corruption according to the docs (although it should work with unique filenames)
    - name: Upload Wheel Artifacts
      if: ${{ env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      uses: actions/upload-artifact@v3
      with:
        name: ${{ env.ARTIFACT_NAME }}
        path: ${{ env.BUILD_DIR }}/lib/${{ env.CONFIG }}/python/dist/*.whl
        if-no-files-found: error
        retention-days: 7


  # Windows Wheel builds
  wheel-windows:
    runs-on: ${{ matrix.cudacxx.os }}
    strategy:
      fail-fast: false
      # Multiplicative build matrix
      # optional exclude: can be partial, include: must be specific
      matrix:
        cudacxx:
          - cuda: "12.0.0"
            cuda_arch: "50-real;60-real;70-real;80-real;90-real;90-virtual"
            hostcxx: "Visual Studio 16 2019"
            os: windows-2019
          - cuda: "11.2.2"
            cuda_arch: "35-real;50-real;60-real;70-real;80-real;80-virtual"
            hostcxx: "Visual Studio 16 2019"
            os: windows-2019
        python: 
          - "3.12"
          - "3.11"
          - "3.10"
          - "3.9"
          - "3.8"
        config:
          - name: "Release"
            config: "Release"
            SEATBELTS: "ON"
        VISUALISATION: 
          - "ON"
          - "OFF"

    # Name the job based on matrix/env options
    name: "wheel-windows (${{ matrix.cudacxx.cuda }}, ${{matrix.python}}, ${{ matrix.VISUALISATION }}, ${{ matrix.config.name }}, ${{ matrix.cudacxx.os }})"
    

    # Define job-wide env constants, and promote matrix elements to env constants for portable steps.
    env:
      # Compute the wheelhouse name which should be unique within the matrix. This must be unique per build matrix/job combination
      ARTIFACT_NAME: wheel-windows-${{ matrix.cudacxx.cuda }}-${{matrix.python}}-${{ matrix.VISUALISATION }}-${{ matrix.config.name }}-${{ matrix.cudacxx.os }}
      # Define constants
      BUILD_DIR: "build"
      FLAMEGPU_BUILD_TESTS: "OFF"
      # Conditional based on matrix via awkward almost ternary
      FLAMEGPU_BUILD_PYTHON: ${{ fromJSON('{true:"ON",false:"OFF"}')[matrix.python != ''] }}
      # Port matrix options to environment, for more portability.
      CUDA: ${{ matrix.cudacxx.cuda }}
      CUDA_ARCH: ${{ matrix.cudacxx.cuda_arch }}
      HOSTCXX: ${{ matrix.cudacxx.hostcxx }}
      OS: ${{ matrix.cudacxx.os }}
      CONFIG: ${{ matrix.config.config }}
      FLAMEGPU_SEATBELTS: ${{ matrix.config.SEATBELTS }}
      PYTHON: ${{ matrix.python}}
      VISUALISATION: ${{ matrix.VISUALISATION }}
      # Ensure MSVC >= 1940 works with CUDA <= 12.3
      CUDAFLAGS: -allow-unsupported-compiler

    steps:
    - uses: actions/checkout@v3

    - name: Install CUDA (Windows)
      if: ${{ runner.os == 'Windows' && env.CUDA != '' }}
      shell: powershell
      env:
        cuda: ${{ env.CUDA }}
        visual_studio: ${{ env.HOSTCXX }}
      run: .github\scripts\install_cuda_windows.ps1

    - name: Select Python
      if: ${{ env.PYTHON != '' && env.FLAMEGPU_BUILD_PYTHON == 'ON' }} 
      uses: actions/setup-python@v4
      with:
        python-version: ${{ env.PYTHON }}

    - name: Install python dependencies
      if: ${{ env.PYTHON != '' && env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      run: |
        python3 -m pip install --upgrade wheel build setuptools

    - name: Add custom problem matchers for annotations
      run: echo "::add-matcher::.github/problem-matchers.json"

    # This pre-emptively patches a bug where ManyLinux didn't generate buildnumber as git dir was owned by diff user
    - name: Enable git safe-directory
      run: git config --global --add safe.directory $GITHUB_WORKSPACE

    # Must pass -G -A for windows, and -DPython3_ROOT_DIR/-DPYTHON3_EXECUTABLE as a github action workaround
    - name: Configure cmake
      run: >
        cmake . -B "${{ env.BUILD_DIR }}" 
        -G "${{ env.HOSTCXX }}" -A x64
        -Werror=dev
        -DCMAKE_WARN_DEPRECATED="OFF"
        -DFLAMEGPU_WARNINGS_AS_ERRORS="ON"
        -DCMAKE_CUDA_ARCHITECTURES="${{ env.CUDA_ARCH }}"
        -DFLAMEGPU_BUILD_TESTS="${{ env.FLAMEGPU_BUILD_TESTS }}"
        -DFLAMEGPU_BUILD_PYTHON="${{ env.FLAMEGPU_BUILD_PYTHON }}" 
        -DPython3_ROOT_DIR="$(dirname $(which python))" 
        -DPython3_EXECUTABLE="$(which python)"
        -DFLAMEGPU_VISUALISATION="${{ env.VISUALISATION }}"
        -DFLAMEGPU_ENABLE_NVTX="ON"

    - name: Build python wheel
      if: ${{ env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      working-directory: ${{ env.BUILD_DIR }}
      run: cmake --build . --config ${{ env.CONFIG }} --target pyflamegpu --verbose -j `nproc`

    # Upload wheel artifacts to the job on GHA, with a short retention
    # Use a unique name per job matrix run, to avoid a risk of corruption according to the docs (although it should work with unique filenames)
    - name: Upload Wheel Artifacts
      if: ${{env.FLAMEGPU_BUILD_PYTHON == 'ON' }}
      uses: actions/upload-artifact@v3
      with:
        name: ${{ env.ARTIFACT_NAME }}
        path: ${{ env.BUILD_DIR }}/lib/${{ env.CONFIG }}/python/dist/*.whl
        if-no-files-found: error
        retention-days: 7


  # Create a draft release, if all other jobs are successful and if the workflow was triggered by a version tag push event.
  create-draft-release:
    needs: 
      - build-ubuntu
      - build-windows
      - wheel-manylinux2014
      - wheel-windows
    if: ${{ success() && startsWith(github.ref, 'refs/tags/v') && github.event_name != 'workflow_dispatch' && github.event_name != 'pull_request' }}
    runs-on: ubuntu-20.04
    steps:
    - uses: actions/checkout@v3

    # Download python wheels from previous jobs.
    - name: Download Wheel Artifacts
      id: download
      uses: actions/download-artifact@v3
      with:
        path: artifacts

    - name: Find Wheel Artifacts
      id: find_wheels
      run: |
        # Find files, storing in the files variable
        pattern="${{ steps.download.outputs.download-path }}/*/*.whl"
        files=""
        for f in ${pattern}; do
          # @future - Set human readable display label
          label="$(basename ${f})"
          files="${files} \"${f}#${label}\""
        done
        echo "files: ${files}"
        # Set the list of files as the output for this step
        echo "files=${files}" >> "$GITHUB_OUTPUT"

    # Extract information from the tag which is required for the draft github release
    - name: Process Tag
      id: tag
      run: |
        ref=${{ github.ref }}
        tag=${ref/refs\/tags\//}
        version=${tag/v/}
        prerelease_label=$(echo ${tag} | cut -d- -f2)
        prerelease_label_len=$(echo ${prerelease_label} | wc -c)
        prerelease_flag=$([[ -z "${prerelease_label_len}" ]] && echo "" || echo "--prerelease")
        # set step outputs
        echo "tag=${tag}" >> "$GITHUB_OUTPUT"
        echo "version=${version}" >> "$GITHUB_OUTPUT"
        echo "prerelease_flag=${prerelease_flag}" >> "$GITHUB_OUTPUT"

    # Use the gh cli tool to create a draft release
    # @future - use --notes "notes string" or --notes-file file
    - name: Create Draft Release
      id: create_release
      run: gh release create --draft ${{ env.PRERELEASE_FLAG}} --title "${{ env.TITLE }}" ${{ env.TAG }} ${{ env.FILES }}
      env:
        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        PRERELEASE_FLAG: "${{ steps.tag.outputs.prerelease_flag }}"
        TAG: "${{ steps.tag.outputs.tag }}"
        TITLE: "FLAME GPU ${{ steps.tag.outputs.version }}"
        FILES: ${{ steps.find_wheels.outputs.files }}