Expand description
Content-addressed JIT code cache with dependency-based invalidation.
Caches compiled native function pointers keyed by FunctionHash.
When the same function blob appears in a subsequent compilation
(same content hash), we skip recompilation and reuse the existing
native code pointer.
Supports dependency tracking: when a function is inlined into another,
the caller records the callee hash as a dependency. If the callee
changes, all dependents can be invalidated via invalidate_by_dependency().
Structsยง
- Cache
Entry - Extended cache entry with dependency tracking.
- JitCode
Cache - Cache of JIT-compiled function pointers, keyed by content hash.