`batuta validate`

Validate semantic equivalence between original and transpiled code (Phase 4).

Synopsis

batuta validate [OPTIONS]

The validate command verifies that transpiled Rust code produces equivalent behavior to the original source. It supports four validation methods: syscall tracing via Renacer, output diffing, test suite execution, and performance benchmarking.

This is Phase 4 of the 5-phase transpilation pipeline. It requires Phase 3 (Optimization) to be completed first.

Options

Option	Description
`--trace-syscalls`	Trace syscalls for comparison using Renacer
`--diff-output`	Compare stdout of original vs transpiled binary
`--run-original-tests`	Run `cargo test` in the transpiled output directory
`--benchmark`	Run performance benchmarks (3 iterations, reports speedup)
`-v, --verbose`	Enable verbose output
`-h, --help`	Print help

Validation Methods

Syscall Tracing (`--trace-syscalls`)

Uses the Renacer syscall tracer to compare system call patterns between the original and transpiled binaries. This provides the deepest semantic equivalence guarantee.

Requires: ./original_binary and ./target/release/transpiled to exist.

Output Diff (`--diff-output`)

Runs both binaries and compares their stdout line-by-line. Shows a unified diff if outputs differ.

Test Execution (`--run-original-tests`)

Runs cargo test in the transpiled output directory (from batuta.toml transpilation.output_dir). Validates that the transpiled code passes its test suite.

Benchmarking (`--benchmark`)

Times both original and transpiled binaries over 3 iterations and reports average execution time and speedup factor.

Examples

Full Validation Suite

$ batuta validate --trace-syscalls --diff-output --run-original-tests --benchmark

✅ Validating equivalence...

Validation Settings:
  • Syscall tracing: enabled
  • Diff output: enabled
  • Original tests: enabled
  • Benchmarks: enabled

🔍 Running Renacer syscall tracing...
  ✅ Syscall traces match - semantic equivalence verified

📊 Output comparison:
  ✅ Outputs match - functional equivalence verified

🧪 Running test suite on transpiled code:
  ✅ All tests pass on transpiled code

⚡ Performance benchmarking:
  Original:   142.3ms avg
  Transpiled:  28.1ms avg
  Speedup:    5.06x faster

Quick Test-Only Validation

$ batuta validate --run-original-tests

Benchmark Comparison

$ batuta validate --benchmark

Exit Behavior

Each validation method independently updates the overall pass/fail status. If any enabled method fails, the Validation phase is marked as failed in the workflow state.

If binaries are not found for --trace-syscalls, --diff-output, or --benchmark, those checks are skipped with a warning (not treated as failures).

The Batuta Book