test.yml

name: Test

on:
  push:
  pull_request:
  workflow_call:

env:
  CARGO_TERM_COLOR: always
  # Enable backtraces for panics but not for "regular" errors.
  RUST_BACKTRACE: 1
  RUST_LIB_BACKTRACE: 0
  RUSTFLAGS: '-D warnings'

jobs:
  build-linux-kernel:
    uses: ./.github/workflows/build-linux.yml
    secrets: inherit
    with:
      # TODO: Bump to v6.11 once tagged.
      rev: 44eacec2cec1c5f83074e74d00208d15390b00bc
      config: 'data/config'
  build:
    name: Build [${{ matrix.runs-on }}, ${{ matrix.rust }}, ${{ matrix.profile }}, ${{ matrix.args }}]
    runs-on: ${{ matrix.runs-on }}
    strategy:
      fail-fast: false
      matrix:
        runs-on: [ubuntu-latest, macos-latest]
        rust: [stable]
        profile: [dev, release]
        # gsym-in-apk requires `generate-unit-test-files` feature, which
        # requires `llvm-gsymutil` installed. We don't want that dependency in
        # this large OS spanning matrix.
        args: ["--workspace --exclude=gsym-in-apk"]
        include:
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            # Make sure to build *without* `--workspace` or feature
            # unification may mean that `--no-default-features` goes
            # without effect.
            args: "--lib --no-default-features"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--lib --no-default-features --features=apk"
          - runs-on: windows-latest
            rust: stable
            profile: release
            args: "--lib --no-default-features"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--lib --no-default-features --features=breakpad"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--lib --no-default-features --features=gsym"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--lib --no-default-features --features=zlib"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--lib --no-default-features --features=zstd"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--package=blazecli"
          - runs-on: ubuntu-latest
            rust: stable
            profile: dev
            args: "--package=blazesym-c"
          - runs-on: ubuntu-latest
            rust: stable
            profile: release
            args: "--package=blazesym-c --no-default-features"
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@master
      with:
        toolchain: ${{ matrix.rust }}
    - uses: Swatinem/rust-cache@v2
      with:
        key: ${{ matrix.runs-on }}-${{ matrix.rust }}-${{ matrix.profile }}
    - name: Build ${{ matrix.profile }}
      run: |
        cargo build --profile=${{ matrix.profile }} ${{ matrix.args }}
  build-cross:
    name: Cross-compile [${{ matrix.target }}]
    runs-on: ubuntu-latest
    strategy:
      fail-fast: false
      matrix:
        target: [
          aarch64-linux-android,
          arm-linux-androideabi,
          armv7-linux-androideabi,
          x86_64-unknown-linux-musl,
        ]
    steps:
      - uses: actions/checkout@v4
      - uses: taiki-e/setup-cross-toolchain-action@v1
        with:
          target: ${{ matrix.target }}
      - run: |
          cargo build --lib
  build-minimum:
    name: Build using minimum versions of dependencies
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: dtolnay/rust-toolchain@nightly
      - run: cargo +nightly -Z minimal-versions update
      - uses: dtolnay/rust-toolchain@master
        with:
          # Please adjust README, `rust-version` field in Cargo.toml,
          # and `msrv` in .clippy.toml when bumping version.
          toolchain: 1.65.0
      - uses: Swatinem/rust-cache@v2
      - run: cargo build --features="apk,backtrace,demangle,dwarf,gsym,tracing"
  nop-rebuilds:
    name: No-op rebuilds
    runs-on: ubuntu-22.04
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@nightly
    - name: Install required tools
      run: sudo apt-get install -y llvm-14
    - name: Check incremental rebuilds
      env:
        CARGO_OPTS: --features=generate-unit-test-files --workspace --quiet --tests
      run: |
        cargo check ${CARGO_OPTS}
        # We need another build here to have the reference `output` file
        # present. As long as we converge eventually it's probably good
        # enough...
        cargo check ${CARGO_OPTS}
        output=$(CARGO_LOG=cargo::core::compiler::fingerprint=info cargo check ${CARGO_OPTS} 2>&1)
        [ -z "${output}" ] || (echo "!!!! cargo check --tests rebuild was not a no-op: ${output} !!!!" && false)
  test-coverage:
    name: Test and coverage
    runs-on: ubuntu-22.04
    needs: [build-linux-kernel]
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@nightly
    - uses: Swatinem/rust-cache@v2
    - uses: actions/setup-python@v5
      id: py312
      with:
        python-version: '3.12'
    - name: Install required tools
      run: sudo apt-get install -y llvm-14
    - name: Install cargo-llvm-cov
      uses: taiki-e/install-action@cargo-llvm-cov
    - name: Install llvm-tools-preview
      # Install llvm-tools-preview, because we don't want this to happen
      # from within our `vmtest` instance.
      run: rustup component add llvm-tools-preview --toolchain nightly-x86_64-unknown-linux-gnu
    - name: Remove .cargo/config.toml
      # We don't want a custom test runner because we run in a VM.
      run: rm .cargo/config.toml
    - name: Build main.sh
      env:
        ARTIFACT_URL: ${{ needs.build-linux-kernel.outputs.kernel-artifact-url }}
        PYTHON: ${{ steps.py312.outputs.python-path }}
      run: |
        # cargo-llvm-cov ultimately is just a thin wrapper around cargo
        # commands with certain environment variables set. Because it
        # does not allow us to build binaries and execute them in
        # different steps, we have to roll our own logic for doing just
        # that, because we do not want to build inside a nested VM...
        cargo llvm-cov show-env | sed 's@.*@export \0@' > env.sh
        echo "export PYTHON=${PYTHON}" >> env.sh
        source env.sh

        cargo test --no-run --workspace --all-targets --features=nightly,generate-large-test-files 2>&1 | tee /tmp/cargo-build.log
        tests=$(IFS='' grep Executable /tmp/cargo-build.log |
          awk '{print $NF}' |
          sed 's@[()]@@g' |
          sed 's@.*@\0 --include-ignored@'
        )
        # Yes, there appears to be no way to just retrieve the
        # uploaded artifact. One can't use actions/download-artifact
        # and provide any of the outputs of actions/upload-artifact.
        # Neither does it seem possible to just download the thing
        # directly, because contents are unconditionally zipped. Good.
        # Lord.
        curl --location \
          --fail-with-body \
          --header "Accept: application/vnd.github+json" \
          --header "Authorization: Bearer ${{ secrets.GITHUB_TOKEN }}"\
          --header "X-GitHub-Api-Version: 2022-11-28" \
          --output artifact.zip \
          "${ARTIFACT_URL}"
        # This unzip will produce the kernel bzImage.
        unzip artifact.zip

        cat <<EOF > main.sh
        #!/bin/sh
        set -e
        . ./env.sh
        $tests
        cargo llvm-cov report --ignore-filename-regex=cli/src/ --lcov --output-path lcov.info
        EOF
        chmod a+x main.sh
    - name: Test and gather coverage
      uses: danobi/vmtest-action@f8c84191b5925c3290595743fc46eaaafd827bf3
      with:
        kernel: bzImage
        command: sh -c 'cd ${{ github.workspace }} && ./main.sh'
    - name: Upload code coverage results
      uses: codecov/codecov-action@v4
      env:
        CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
      with:
        files: lcov.info
  build-test-artifacts:
    # Only run Windows tests on the final push. They are very unlikely to break
    # at runtime (we build them anyway), but take quite a bit longer than Linux
    # based ones.
    if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/main' }}
    name: Build test artifacts for Windows
    runs-on: ubuntu-22.04
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - name: Install required tools
      run: sudo apt-get install -y llvm-14
    - uses: dtolnay/rust-toolchain@stable
    - uses: Swatinem/rust-cache@v2
    - run: cargo build --features=generate-unit-test-files
    - uses: actions/upload-artifact@v4
      with:
        name: test-stable-addrs
        if-no-files-found: error
        path: |
          data/test-stable-addrs*
          data/test-rs.bin
  test-windows:
    name: Test on Windows
    runs-on: windows-latest
    needs: [build-test-artifacts]
    steps:
    - uses: actions/checkout@v4
    - uses: actions/download-artifact@v4
      with:
        name: test-stable-addrs
        path: data/
    - uses: dtolnay/rust-toolchain@stable
    - uses: Swatinem/rust-cache@v2
    - run: cargo test --tests --release --features=dont-generate-unit-test-files -- ':windows:'
  test-sanitizers:
    name: Test with ${{ matrix.sanitizer }} sanitizer
    strategy:
      fail-fast: false
      matrix:
        sanitizer: [address, leak]
    runs-on: ubuntu-latest
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@nightly
    - uses: Swatinem/rust-cache@v2
      with:
        key: ${{ matrix.sanitizer }}
    - name: Enable debug symbols
      run: |
          # to get the symbolizer for debug symbol resolution
          sudo apt-get install -y llvm-14
          # to fix buggy leak analyzer:
          # https://github.com/japaric/rust-san#unrealiable-leaksanitizer
          sed -i '/\[features\]/i [profile.dev]' Cargo.toml
          sed -i '/profile.dev/a opt-level = 1' Cargo.toml
          cat Cargo.toml
    - name: cargo test -Zsanitizer=${{ matrix.sanitizer }}
      env:
        # TODO: Ideally we'd set CFLAGS and CXXFLAGS as well, but some
        #       of our transitive dependencies don't play nice with
        #       ASAN.
        #CFLAGS: "-fsanitize=${{ matrix.sanitizer }}"
        #CXXFLAGS: "-fsanitize=${{ matrix.sanitizer }}"
        RUSTFLAGS: "-Zsanitizer=${{ matrix.sanitizer }}"
        ASAN_OPTIONS: "detect_odr_violation=0:detect_leaks=0"
        LSAN_OPTIONS: ""
      run: cargo test --workspace --lib --tests --target x86_64-unknown-linux-gnu
  test-release:
    name: Test with release build
    runs-on: ubuntu-22.04
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - name: Install required tools
      run: sudo apt-get install -y llvm-14
    - uses: dtolnay/rust-toolchain@stable
    - uses: Swatinem/rust-cache@v2
    - run: cargo test --workspace --release
  test-miri:
    name: Test with Miri
    runs-on: ubuntu-latest
    env:
      # Stack borrows are very experimental. Disable for now.
      MIRIFLAGS: '-Zmiri-disable-stacked-borrows'
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@master
      with:
        # TODO: Recent nightly versions fail with what *seems* to be a
        #       false positive. Fall back to a known good version. We
        #       should come back to this eventually.
        toolchain: nightly-2024-08-06
        components: miri
    # Miri would honor our custom test runner, but doesn't work with it. We
    # could conceivably override that by specifying
    # CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_RUNNER, except it appears as if Miri
    # uses the runner itself. In short, it's a mess. Just remove any
    # such custom configuration when running Miri.
    - name: Remove .cargo/config.toml
      run: rm .cargo/config.toml
    - name: Run tests
      run: cargo miri test --workspace --features=dont-generate-unit-test-files -- ":miri:"
  test-examples:
    name: Test examples
    runs-on: ubuntu-22.04
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@stable
    - run: cargo run --example backtrace
    - run: cargo run --package gsym-in-apk
  c-header:
    name: Check generated C header
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@stable
    - run: cargo check --package=blazesym-c --features=generate-c-header
    - name: Check that C header is up-to-date
      run: git diff --exit-code ||
             (echo "!!!! CHECKED IN C HEADER IS OUTDATED !!!!" && false)
  bench:
    # Only run benchmarks on the final push. They are generally only
    # informative because the GitHub Runners do not provide a stable
    # performance baseline anyway.
    if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/main' }}
    name: Benchmark
    runs-on: ubuntu-22.04
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
    steps:
    - uses: actions/checkout@v4
    - uses: dtolnay/rust-toolchain@nightly
    - name: Install required tools
      run: sudo apt-get install -y llvm-14
    - uses: Swatinem/rust-cache@v2
    - name: Run benchmarks
      shell: bash
      run: |
        echo '```' >> $GITHUB_STEP_SUMMARY
        cargo bench --workspace --features=nightly -- bench_ | tee --append $GITHUB_STEP_SUMMARY
        # We use bencher format here for better relation to the above
        # but also because it emits less other crap into our summary.
        # Note that because libtest does not understand the
        # `--output-format` option, we need to specify the benchmark
        # binary (`main`) here and have a different invocation for
        # libtest style benchmarks above. Sigh.
        cargo bench --workspace --bench=main --bench=capi --features=generate-large-test-files,dont-generate-unit-test-files -- --output-format=bencher | tee --append $GITHUB_STEP_SUMMARY
        echo '```' >> $GITHUB_STEP_SUMMARY
  clippy:
    name: Lint with clippy
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: dtolnay/rust-toolchain@stable
      - run: cargo clippy --workspace --no-deps --all-targets --features=dont-generate-unit-test-files -- -A unknown_lints -D clippy::todo
  rustfmt:
    name: Check code formatting
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: dtolnay/rust-toolchain@nightly
        with:
          components: rustfmt
      - run: cargo +nightly fmt --all -- --check
  cargo-doc:
    name: Generate documentation
    runs-on: ubuntu-22.04
    env:
      LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-14
      RUSTDOCFLAGS: '--cfg docsrs -D warnings'
    steps:
      - uses: actions/checkout@v4
      - uses: dtolnay/rust-toolchain@nightly
      - run: cargo doc --workspace --no-deps