Test

Abstract away more ElfN_Ehdr attributes #2416

Workflow file for this run

	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]
	rust: [stable]
	profile: [dev, release]
	# gsym-in-apk requires `blazesym-dev/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=blazesym-dev --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: windows-latest
	rust: stable
	profile: dev
	args: "--tests --features=blazesym-dev/dont-generate-unit-test-files"
	- runs-on: macos-latest
	rust: stable
	profile: release
	args: "--lib --no-default-features"
	- runs-on: macos-latest
	rust: stable
	profile: dev
	args: "--tests --features=blazesym-dev/dont-generate-unit-test-files"
	- 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=bpf"
	- 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:
	# Please update the build-cli workflow cross-compile matrix when
	# adding a new target here.
	target: [
	aarch64-linux-android,
	arm-linux-androideabi,
	armv7-linux-androideabi,
	i686-unknown-linux-gnu,
	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
	cargo build --lib --features=bpf
	cargo build --package=blazesym-c --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: 1.69.0
	- uses: Swatinem/rust-cache@v2
	- run: cargo build --features="apk,backtrace,bpf,demangle,dwarf,gsym,tracing"
	nop-rebuilds:
	name: No-op rebuilds
	runs-on: ubuntu-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@nightly
	- name: Check incremental rebuilds
	env:
	CARGO_OPTS: --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-24.04
	needs: [build-linux-kernel]
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev
	- 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 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: ${{ steps.py312.outputs.python-path }}
	RUSTFLAGS: '--cfg has_large_test_files'
	run: \|
	# cargo-llvm-cov ultimately is just a thin wrapper around cargo
	# commands with certain environment variables set. Because it
	# does not allow us to build binaries and execute them in
	# different steps, we have to roll our own logic for doing just
	# that, because we do not want to build inside a nested VM...
	cargo llvm-cov show-env \| sed 's@.*@export \0@' > env.sh
	echo "export PYTHON=${PYTHON}" >> env.sh
	source env.sh

	cargo test --no-run --workspace --all-targets --features=nightly,blazesym-dev/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@v5
	env:
	CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
	with:
	files: lcov.info
	build-test-artifacts:
	# Only run tests on other operating systems 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
	runs-on: ubuntu-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@stable
	- run: cargo check --package=blazesym-dev --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
	data/libtest-so*
	test-partly-supported:
	name: Test on ${{ matrix.runs-on }}
	strategy:
	fail-fast: false
	matrix:
	runs-on: [macos-latest, windows-latest]
	runs-on: ${{ matrix.runs-on }}
	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
	- run: cargo test --tests --release --features=blazesym-dev/dont-generate-unit-test-files -- ':other_os:'
	test-sanitizers:
	name: Test with ${{ matrix.sanitizer }} sanitizer
	strategy:
	fail-fast: false
	matrix:
	sanitizer: [address]
	runs-on: ubuntu-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@nightly
	- uses: Swatinem/rust-cache@v2
	with:
	key: ${{ matrix.sanitizer }}
	- name: cargo test -Zsanitizer=${{ 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=1"
	run: cargo test --workspace --lib --tests --target x86_64-unknown-linux-gnu
	test-release:
	name: Test with release build
	runs-on: ubuntu-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev liblzma-dev
	- uses: actions/checkout@v4
	- 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:
	- name: Install development dependencies
	run: sudo apt-get install -y libelf-dev zlib1g-dev
	- 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=blazesym-dev/dont-generate-unit-test-files -- ":miri:"
	test-examples:
	name: Test examples
	runs-on: ubuntu-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@stable
	- run: cargo run --example backtrace
	- run: cargo run --example inspect-mangled
	- run: cargo run --example normalize-virt-offset
	- 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-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	steps:
	- name: Install development dependencies
	run: sudo apt-get install -y gcc-multilib libelf-dev zlib1g-dev
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@nightly
	- 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=blazesym-dev/generate-large-test-files,blazesym-dev/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:
	- name: Install development dependencies
	run: sudo apt-get install -y libelf-dev zlib1g-dev
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@stable
	- run: cargo clippy --workspace --no-deps --all-targets --features=blazesym-dev/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-24.04
	env:
	LLVM_GSYMUTIL: /usr/bin/llvm-gsymutil-18
	RUSTDOCFLAGS: '--cfg docsrs -D warnings'
	steps:
	- uses: actions/checkout@v4
	- uses: dtolnay/rust-toolchain@nightly
	- run: cargo doc --workspace --exclude=blazesym-dev --exclude=gsym-in-apk --no-deps

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Abstract away more ElfN_Ehdr attributes #2416

Workflow file

Abstract away more ElfN_Ehdr attributes #2416

Jobs

Run details

Workflow file for this run