Update 2025-11-NN-swift-gsoc-2025-highlight-4-swift-testing-output.md

Co-authored-by: Joseph Heck <j_heck@apple.com>

Author: ktoso

_posts/2025-11-NN-swift-gsoc-2025-highlight-4-swift-testing-output.md

@@ -78,7 +78,7 @@ The new recorder introduces a rich, hierarchical display that immediately brings
 
 * Designing for serialization & ABI: Refactoring to consume ABI.EncodedEvent taught me the value of decoupling - moving from in-process object graphs to a serialized event stream increases robustness and enables harness-style execution. It also exposed careful trade-offs: serialization simplifies process boundaries but requires stable encoding contracts and more explicit discovery/lookup logic for test metadata.
 * Building a thread-safe collector for results required balancing lock granularity and throughput. The adopted strategy (a single lock per event bucket, plus careful mutation patterns) simplifies correctness while keeping contention low for typical test workloads.
-* Cross-platform fallbacks: Supporting ASCII fallback for line-drawing characters taught me to design for the weakest terminal capability first, then progressively enhance for modern terminals. Testing across Windows/Linux/macOS terminals and non-interactive CI runners uncovered many small but critical edge-cases (line-wrapping, width detection fallbacks).
+* Cross-platform fallbacks: Supporting ASCII fallback for line-drawing characters taught me to design for the weakest terminal capability first, then progressively enhance for modern terminals. Testing across Windows/Linux/macOS terminals and non-interactive CI runners uncovered many small but critical edge cases (line-wrapping, width detection fallbacks).
 * Deciding what to show inline vs. in a details section is a UX trade-off; too much inline detail makes the tree noisy, and too little hides useful context. The two-tier approach (short inline message + full details section) balances these needs.
 
 **Collaboration & process lessons:**