Debugging with Valgrind + Nextest

When you encounter crashes, segfaults, or memory errors in tests, use valgrind for proper debugging.

Quick Usage

# Run a specific test under valgrind
cargo nextest run --profile valgrind -p PACKAGE --test TEST_FILE TEST_NAME

# Example
cargo nextest run --profile valgrind -p facet-format-json --test jit_deserialize test_jit_option_some

How It Works

The project has valgrind pre-configured in .config/nextest.toml:

1. Wrapper script: Defined as [scripts.wrapper.valgrind]

   • Command: valgrind --leak-check=full --show-leak-kinds=all ...
   • Automatically wraps test execution

2. Profile: [profile.valgrind] applies the wrapper to all tests on Linux

3. Running: Use --profile valgrind flag with nextest

Benefits

• ✅ Automatic setup - no manual valgrind commands needed
• ✅ Proper configuration - leak checking, error codes pre-configured
• ✅ Integrated - works with nextest filtering and test selection
• ✅ Clean output - nextest captures and formats valgrind output

Don't Do This

❌ Running valgrind manually: valgrind ./target/debug/deps/test-binary ❌ Using raw --no-run + valgrind commands ❌ Playing hunt-the-segfault without proper tools

Do This Instead

✅ Use the pre-configured profile: cargo nextest run --profile valgrind <filters> ✅ Let nextest handle the wrapper script integration ✅ Get clean, actionable valgrind output immediately

Debugging Workflow

1. Test crashes with SIGSEGV
2. Run: cargo nextest run --profile valgrind -p PACKAGE --test FILE TEST_NAME
3. Valgrind shows exact line where invalid read/write occurs
4. Fix the bug
5. Verify with regular tests

See Also

• Nextest wrapper scripts docs: https://nexte.st/docs/configuration/wrapper-scripts/
• Project nextest config: .config/nextest.toml