Scaling Static Code Analysis Adoption at WhatsApp iOS
summary
WhatsApp iOS deploys Infer and Clang Static Analyzer across diffs and daily CI, reporting ~800 Infer issues (700 fixed, 86% fix rate) and ~2,400 Clang issues (≈1,000 fixed initially, 42% rising to ~70% after bisection relint). The effort prevents over 1700 regressions, mobilizes 50+ engineers, uses IDE‑friendly incremental checks, gamified cleanup events that produced ~400 Infer fixes, and experiments with generative AI that remediated 1000+ legacy issues but sometimes introduces hallucinated bugs.
AI-written summary · about
chapters (10)
- 0:00Scaling static analysis at WhatsApp iOS
- 1:28Selecting Infer and Clang Static Analyzer
- 4:38Deployment workflow for static analysis
- 6:48Stopping regressions with Infer
- 8:14Challenges with Clang Tidy
- 11:06Cleaning up pre-existing issues
- 13:15Improving issue tracking UX
- 14:39Engineer feedback and maintenance costs
- 15:42AI and GenAI for static analysis
- 17:08Q&A
related talks
27mSupercharging Infer: Enhancing Static Analysis with Dynamic Data at WhatsApp
45mAnalyze This
31mExploring Static Code Analysis techniques in Flutter
52mLayers of testing in iOS applications
26mStatic Code Analysis with Gitlab CI
01h 00mCode Quality in Practice
24mWriting Stable iOS Applications
55mAutomated Testing in an iOS Project: The ICQ Case