Semantic Subsystems

Semantic subsystems are used by several compilation phases and therefore do not form a simple linear pipeline. Name resolution participates in expansion, signature checking, and body checking. Type inference and trait solving cooperate during body checking. External metadata supplies authoritative facts about dependency definitions without deciding Sage's semantic questions.

Each subsystem chapter explains its public semantic boundary, permitted dependencies, important algorithms, and the phases that invoke it. Claims beside the modeled entities identify the load-bearing invariants. Their generated anchor pages display implementation entries, tests, saved reviews, and explicit TODO/N/A gaps; these sources supply coverage without a separate subsystem status table.

• Name Resolution maps scoped Rust syntax to symbols and preserves ambiguity or incompleteness.
• Type Inference owns body-local constraints, speculative versions, obligations, and finalization.
• Trait Solver proves canonical propositions and produces goal-specific outputs such as normalized types.
• External Metadata imports narrow authoritative dependency facts without delegating Sage semantic decisions.