datahaven/.github/workflows/actions/setup-env/action.yml

name: "Setup Rust Environment"
description: "Creates a Rust environment with the specified toolchain, cache, and dependencies"

inputs:
  cache-key:
    description: "Cache key used to retrieve built data. Usually matches the profile of the build"
    required: false
    default: "cache"
  
  install-deps:
    description: "Whether to install system dependencies. Set to false for self-hosted runners with pre-installed deps"
    required: false
    default: "true"

runs:
  using: "composite"
  steps:
    - name: Set tool versions and cache key
      shell: bash
      run: |
        # Define tool versions in one place
        echo "MOLD_VERSION=2.40.4" >> $GITHUB_ENV
        echo "LLVM_VERSION=18.1.8" >> $GITHUB_ENV
        echo "PROTOC_VERSION=28.3" >> $GITHUB_ENV
        echo "LIBPQ_VERSION=18.1-1" >> $GITHUB_ENV

        # Create a hash from tool versions for cache key
        TOOLS_HASH=$(echo "mold-2.40.4|llvm-18.1.8|protoc-28.3|libpq-18.1-1" | sha256sum | cut -c1-16)
        echo "TOOLS_HASH=$TOOLS_HASH" >> $GITHUB_ENV
        echo "Tools cache key hash: $TOOLS_HASH"        

    - name: Set Rust version
      shell: bash
      run: |
        echo "BUILD_RUST_VERSION=$(rustc --version)" >> $GITHUB_ENV        

    - name: Run sccache-cache
      uses: mozilla-actions/sccache-action@v0.0.9

    - name: Show sccache stats
      shell: bash
      run: sccache --show-stats

    # Cache specific ~/.local subdirs for locally installed tools (mold, llvm, protoc, libpq)
    # Note: We don't cache all of ~/.local to avoid container storage in ~/.local/share
    - name: Cache local tools
      uses: actions/cache@v4
      with:
        path: |
          ~/.local/bin
          ~/.local/lib
          ~/.local/include          
        key: ${{ runner.os }}-local-tools-${{ env.TOOLS_HASH }}

    # Set up PATH and env vars for cached tools (must run after cache restore)
    - name: Setup paths for cached tools
      shell: bash
      run: |
        # Add ~/.local/bin to PATH for cached tools
        echo "$HOME/.local/bin" >> $GITHUB_PATH
        export PATH="$HOME/.local/bin:$PATH"

        # Setup library paths if ~/.local/lib exists (from cached libpq/llvm)
        if [ -d "$HOME/.local/lib" ]; then
          echo "LD_LIBRARY_PATH=$HOME/.local/lib:${LD_LIBRARY_PATH:-}" >> $GITHUB_ENV
          echo "LIBRARY_PATH=$HOME/.local/lib:${LIBRARY_PATH:-}" >> $GITHUB_ENV
          echo "PKG_CONFIG_PATH=$HOME/.local/lib/pkgconfig:${PKG_CONFIG_PATH:-}" >> $GITHUB_ENV
          echo "LDFLAGS=-L$HOME/.local/lib ${LDFLAGS:-}" >> $GITHUB_ENV
          echo "RUSTFLAGS=-L $HOME/.local/lib" >> $GITHUB_ENV
        fi

        # Setup include paths if ~/.local/include exists
        if [ -d "$HOME/.local/include" ]; then
          echo "C_INCLUDE_PATH=$HOME/.local/include:${C_INCLUDE_PATH:-}" >> $GITHUB_ENV
          echo "CPLUS_INCLUDE_PATH=$HOME/.local/include:${CPLUS_INCLUDE_PATH:-}" >> $GITHUB_ENV
          echo "CPPFLAGS=-I$HOME/.local/include ${CPPFLAGS:-}" >> $GITHUB_ENV
        fi

        # Setup LLVM paths if libclang exists in cache
        if [ -d "$HOME/.local/lib" ] && ls $HOME/.local/lib/libclang* &>/dev/null; then
          echo "LIBCLANG_PATH=$HOME/.local/lib" >> $GITHUB_ENV
          echo "LLVM_CONFIG_PATH=$HOME/.local/bin/llvm-config" >> $GITHUB_ENV
        fi

        # Export mold path if it exists
        if command -v mold &>/dev/null; then
          echo "MOLD_PATH=$(which mold)" >> $GITHUB_ENV
        fi        

    - uses: actions/cache@v4
      with:
        path: |
          ~/.cargo/bin/
          ~/.cargo/registry/index/
          ~/.cargo/registry/cache/
          ~/.cargo/git/db/          
        key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
    
    - name: Setup Rust toolchain
      uses: dtolnay/rust-toolchain@stable
      with:
        components: rustfmt, clippy
    
    # Install mold - always install locally to avoid sudo requirements
    - name: Install mold
      shell: bash
      run: |
        # Check if mold is already available and the right version
        if ! command -v mold &> /dev/null || [[ $(mold --version 2>/dev/null | grep -oP '\d+\.\d+\.\d+' | head -1) != "$MOLD_VERSION" ]]; then
          echo "Installing mold $MOLD_VERSION in ~/.local"
          mkdir -p ~/.local/bin
          wget -q -O- "https://github.com/rui314/mold/releases/download/v${MOLD_VERSION}/mold-${MOLD_VERSION}-$(uname -m)-linux.tar.gz" | \
            tar -xzf - -C ~/.local --strip-components=1
          echo "$HOME/.local/bin" >> $GITHUB_PATH
          export PATH="$HOME/.local/bin:$PATH"
        else
          echo "mold is already installed: $(mold --version)"
        fi

        # Export the mold path for use in RUSTFLAGS
        MOLD_PATH=$(which mold)
        echo "MOLD_PATH=${MOLD_PATH}" >> $GITHUB_ENV
        echo "Mold installed at: ${MOLD_PATH}"        
    
    # Install system dependencies only when install-deps is true
    - name: Install system dependencies
      if: inputs.install-deps == 'true'
      shell: bash
      run: sudo apt-get update && sudo apt-get install -y libpq-dev libclang-dev
    
    # Auto-install missing dependencies when install-deps is false (for self-hosted runners)
    # Note: PATH and env vars are already set up in "Setup paths for cached tools" step
    - name: Setup system dependencies (self-hosted)
      if: inputs.install-deps == 'false'
      shell: bash
      run: |
        echo "Checking and installing system dependencies locally if needed..."

        # Setup local directories
        LOCAL_DIR="$HOME/.local"
        LOCAL_BIN="$LOCAL_DIR/bin"
        LOCAL_LIB="$LOCAL_DIR/lib"
        LOCAL_INCLUDE="$LOCAL_DIR/include"
        LOCAL_PKG_CONFIG="$LOCAL_DIR/lib/pkgconfig"

        mkdir -p "$LOCAL_BIN" "$LOCAL_LIB" "$LOCAL_INCLUDE" "$LOCAL_PKG_CONFIG"

        # Check and install PostgreSQL client libraries if needed
        if ! pkg-config --exists libpq 2>/dev/null; then
          echo "Installing libpq $LIBPQ_VERSION..."
          ARCH=$(dpkg --print-architecture 2>/dev/null || echo "amd64")
          WORK_DIR="$LOCAL_DIR/tmp"
          mkdir -p "$WORK_DIR" && cd "$WORK_DIR"

          # Download PostgreSQL 18 packages
          wget -q "https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-18/libpq5_${LIBPQ_VERSION}.pgdg+2_${ARCH}.deb"
          wget -q "https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-18/libpq-dev_${LIBPQ_VERSION}.pgdg+2_${ARCH}.deb"

          # Extract and install
          for pkg in *.deb; do
            ar p "$pkg" data.tar.xz | tar -xJ --no-same-owner 2>/dev/null || true
          done

          cp -r usr/lib/*/libpq* usr/lib/libpq* "$LOCAL_LIB/" 2>/dev/null || true
          cp -r usr/include/* "$LOCAL_INCLUDE/" 2>/dev/null || true
          find usr -name "*.pc" -exec cp {} "$LOCAL_PKG_CONFIG/" \; 2>/dev/null || true

          # Fix pkg-config paths
          sed -i "s|/usr|$LOCAL_DIR|g" "$LOCAL_PKG_CONFIG"/*.pc 2>/dev/null || true

          cd "$HOME" && rm -rf "$WORK_DIR"

          echo "✓ libpq installed"
        else
          echo "✓ libpq found: $(pkg-config --modversion libpq 2>/dev/null || echo 'version unknown')"
        fi

        # Check and install LLVM/Clang if needed
        if ! command -v clang &> /dev/null || ! ls $LOCAL_LIB/libclang* 2>/dev/null && ! ldconfig -p 2>/dev/null | grep -q libclang; then
          echo "LLVM/Clang not found, installing locally..."
          ARCH=$(uname -m)

          # Map architecture names
          case "$ARCH" in
            x86_64) LLVM_ARCH="x86_64-linux-gnu-ubuntu-22.04" ;;
            aarch64) LLVM_ARCH="aarch64-linux-gnu" ;;
            *) LLVM_ARCH="$ARCH-linux-gnu" ;;
          esac

          # Download pre-built LLVM/Clang binaries
          echo "Downloading LLVM ${LLVM_VERSION} for ${LLVM_ARCH}..."
          wget -q -O- "https://github.com/llvm/llvm-project/releases/download/llvmorg-${LLVM_VERSION}/clang+llvm-${LLVM_VERSION}-${LLVM_ARCH}.tar.xz" | \
            tar -xJf - -C /tmp --strip-components=1 || {
              echo "Failed to download pre-built binaries, trying alternative version..."
              # Try an older, more widely available version
              LLVM_FALLBACK="17.0.6"
              wget -q -O- "https://github.com/llvm/llvm-project/releases/download/llvmorg-${LLVM_FALLBACK}/clang+llvm-${LLVM_FALLBACK}-${LLVM_ARCH}.tar.xz" | \
                tar -xJf - -C /tmp --strip-components=1
          }

          # Copy required files
          if [ -d "/tmp/bin" ]; then
            cp /tmp/bin/clang* "$LOCAL_BIN/" 2>/dev/null || true
            cp /tmp/bin/llvm-config "$LOCAL_BIN/" 2>/dev/null || true
            cp -r /tmp/lib/libclang* "$LOCAL_LIB/" 2>/dev/null || true
            cp -r /tmp/lib/clang "$LOCAL_LIB/" 2>/dev/null || true
            cp -r /tmp/include/clang* "$LOCAL_INCLUDE/" 2>/dev/null || true
            rm -rf /tmp/bin /tmp/lib /tmp/include /tmp/share

            # Set LIBCLANG_PATH for rust bindgen (in case not set by cache)
            echo "LIBCLANG_PATH=$LOCAL_LIB" >> $GITHUB_ENV
            echo "LLVM_CONFIG_PATH=$LOCAL_BIN/llvm-config" >> $GITHUB_ENV
          fi

          echo "✓ LLVM/Clang ${LLVM_VERSION} installed locally"
        else
          echo "✓ clang found: $(clang --version 2>/dev/null | head -n1 || echo 'installed')"
        fi

        echo "All required system dependencies ready!"        
    
    # Install protoc only when install-deps is true
    - name: Install Protoc
      if: inputs.install-deps == 'true'
      uses: arduino/setup-protoc@v3
      with:
        repo-token: ${{ github.token }}
    
    # Auto-install protoc when install-deps is false (for self-hosted runners)
    - name: Setup Protoc (self-hosted)
      if: inputs.install-deps == 'false'
      shell: bash
      run: |
        echo "Checking and installing protoc locally if needed..."

        if ! command -v protoc &> /dev/null; then
          echo "protoc not found, installing locally..."

          ARCH=$(uname -m)
          
          # Map architecture names for protoc
          case "$ARCH" in
            x86_64) PROTOC_ARCH="x86_64" ;;
            aarch64) PROTOC_ARCH="aarch_64" ;;
            *) PROTOC_ARCH="$ARCH" ;;
          esac
          
          # Download and install protoc
          mkdir -p ~/.local/bin
          wget -q -O /tmp/protoc.zip "https://github.com/protocolbuffers/protobuf/releases/download/v${PROTOC_VERSION}/protoc-${PROTOC_VERSION}-linux-${PROTOC_ARCH}.zip"
          unzip -q -o /tmp/protoc.zip -d ~/.local
          rm /tmp/protoc.zip
          
          # Make sure it's executable
          chmod +x ~/.local/bin/protoc
          
          # Add to PATH if not already there
          echo "$HOME/.local/bin" >> $GITHUB_PATH
          export PATH="$HOME/.local/bin:$PATH"
          
          echo "✓ protoc ${PROTOC_VERSION} installed locally"
        else
          PROTOC_VERSION=$(protoc --version | grep -oP '\d+\.\d+(\.\d+)?')
          echo "✓ protoc found: version $PROTOC_VERSION"
          
          # Check minimum version (3.0 or higher recommended)
          MAJOR_VERSION=$(echo $PROTOC_VERSION | cut -d. -f1)
          if [ "$MAJOR_VERSION" -lt 3 ]; then
            echo "WARNING: protoc version $PROTOC_VERSION is older than recommended (3.0+)"
          fi
        fi
        
        echo "Protoc ready!"