From #58 (Template Inheritance - Phase 5) code review. Optimize resolver.rs:916-927 to fold frontmatter accumulation once at the leaf instead of re-accumulating transitive FM at every intermediate base in an @extends chain. The condition is bounded (depth-capped) so currently acceptable; now guarded by performance test p5b_deep_chain_32_levels_with_frontmatter_under_2s (commit 14e5d6e). Structural refactoring to eliminate algorithmic inefficiency.
From #58 (Template Inheritance - Phase 5) code review. Optimize resolver.rs:916-927 to fold frontmatter accumulation once at the leaf instead of re-accumulating transitive FM at every intermediate base in an @extends chain. The condition is bounded (depth-capped) so currently acceptable; now guarded by performance test p5b_deep_chain_32_levels_with_frontmatter_under_2s (commit 14e5d6e). Structural refactoring to eliminate algorithmic inefficiency.