JRuby: Insights from Six Years in Production
summary
JRuby asserts concrete differences from MRI: native threads on the JVM, tuned GC, and upcoming Invoke Dynamic speedups. A parallel benchmark runs FIB(1000) 100,000 times and shows 41s → 1.4s (CPU ~10s) after JIT warmup. Java libraries (itext, Swing), Spring integration and wrapping JRuby objects as Java beans are demonstrated, with deployment via vendored gems and WAR/Jetty. Limitations: no fork no continuations. Several migrations work technically; a bottom‑up port is feasible but incremental silo rewrites deliver the business wins.
AI-written summary · about
chapters (7)
- 0:00Start
- 5:33What's different?
- 9:19Parallelism in JRuby
- 18:41Java Integration
- 26:36Case Study 1
- 27:26Case Study 2
- 33:40Case Study 3
