benchmark

name	benchmark
description	Run and analyze performance benchmarks. Use when working on performance-critical code or comparing changes.

Performance Benchmarking

Run the Criterion benchmarks and analyze performance.

Running Benchmarks

# Run the whole suite
cargo bench

# Run a single benchmark by name filter
cargo bench --bench eval -- parse
cargo bench --bench eval -- eval_str/field_access
cargo bench --bench eval -- eval_str/iterate_100

Comparing Against Baseline

# Save a baseline before changes
cargo bench --bench eval -- --save-baseline before

# Make your changes...

# Compare against the baseline
cargo bench --bench eval -- --baseline before

Benchmarks

Benchmark group	File	What it measures
parse/*	benches/eval.rs	Compiling a filter (parser::parse)
eval_str/*	benches/eval.rs	End-to-end eval_str (parse filter + load YAML + evaluate)

Performance Guidelines

• No regressions in hot paths (>5% slower requires justification)
• Keep filter parsing allocation-light
• Prefer iterating over buffering when handling large documents

When to Use

• Before merging performance-sensitive changes
• When optimizing the lexer / parser / evaluator
• To establish a baseline before refactoring