Leveraging Composable Architecture at Scale
summary
Shows concrete fixes and measured outcomes: switching from a single global store to multi-store cuts typing CPU from 9s to 0.9s for a simple case, and noop actions cost falls from ~6ms to ~2.3–3ms after optimizations. Recommends turning off exhaustive testing for non-leaf reducers to avoid huge, fragile tests and using a FeatureAction protocol (view/delegate/local) plus SwiftLint to enforce boundaries. Practical tools include a performance dashboard, a filtered spy of actions, and a lifetime tracker that finds store leaks.
AI-written summary · about
chapters (7)
- 0:00Introduction and TCA overview
- 5:26Testing challenges in TCA
- 11:25Maintainability and feature boundaries
- 16:50Performance pitfalls in TCA
- 21:26Multi-store vs single-store architecture
- 27:29Tooling for large-scale TCA projects
- 31:45Summary and recommendations
related talks
40mTCA: Path towards Building Better SwiftUI Apps
30mOptimizing for Developer Delight
54mPerformance Architecture
23mReducing App Launch Times: From Diagnostics to Optimization
27mDeveloper Experience: It’s Time to Start Complaining
35mSwiftUI Performance for Demanding Apps
35mProgramming without recompiling
29mSpeeding up as an iOS developer