Speaking in Different Tongues: balancing a code base across different languages
summary
Open source Java PDF library is kept functionally identical in C# by an automated conversion pipeline built around a heavily modified sharpen tool that produces an autoport output, runs in Jenkins, enforces Sonar quality checks and unit tests, and blocks merges when conversion touches a curated list of manual files. Porting cuts cross‑platform release lag from weeks to near‑instant and works reliably, with manual handling for crypto, generics and lambdas and a plan to move the manual‑file list into each project.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:16Who am I
- 1:27What is A Price
- 2:00What is Itex
- 3:51Multiplatforms
- 4:13functionally equivalent
- 5:55old times
- 7:54painful process
- 8:33Sharpen
- 9:00Manual porting
- 10:30Autoport
- 11:10Build strategy
- 12:23Future plans
related talks
27mLGTM: Prepping a Mature Open-Source Project for GraalVM
26mAutomating Language Migration at Scale: Tools, Tradeoffs, and the Role of AI
46mOpen sourcing a library: how hard can that be?
49mOpen sourcing a library: how hard can that be?
28mCode Merging Using Transformations and Member Identity
53mPorting Code - Lessons Learned
46mTackling Language Barriers in Open Source Docs: A Case Study of openEuler's...
33mToolchains Are Only As Strong As Their Weakest Link: A Discussion On Interoperability Across CI/C...