PerryTS/perry v0.4.14

on GitHub

Performance

• Native fcmp for numeric comparisons: known-numeric operands emit Cranelift fcmp instead of js_jsvalue_compare runtime function call — mandelbrot 30% faster
• compile_condition_to_bool fast path: numeric Compare in loop/if conditions produces I8 boolean directly, skipping NaN-box round-trip
• Short-circuit &&/||: proper branching in compile_condition_to_bool instead of always-evaluate-both with band/bor
• In-place string append with capacity tracking: js_string_append reuses allocation when refcount=1 and capacity allows — string_concat benchmark 125x faster (900ms → 7ms)
• Deferred module-var write-back in loops: skip global stores inside simple loops (no function calls), flush at exit
• Method inlining for small class methods

Bug Fixes

• Linux linker no longer requires PulseAudio for non-UI programs — -lpulse-simple -lpulse moved behind needs_ui guard (GH-8)
• perry run . now works — positional args parsed flexibly so non-platform values are treated as input path instead of erroring

Benchmarks

Full benchmark overhaul — all benchmarks rerun with Node.js v25 and Bun 1.3, with Bun included in every entry. README now shows the full picture with context explaining both wins and losses:

V8 leads on f64-intensive workloads due to SIMD auto-vectorization and decades of JIT optimization. Perry excels at function calls, recursion, and sequential memory access.