You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
perf: skip cache and deep_freeze for non-LE templates
When a template has no AWS::LanguageExtensions transform, the work
being cached is a single dict lookup on template.get('Transform') —
O(1). The previous code still deep_freeze'd the entire template O(n)
and stored it in the module-level cache, which was pure overhead for
the vast majority of users who don't use language extensions.
Now the no-LE path returns a LanguageExtensionResult wrapping the
caller's original dict directly — zero copies, zero cache entries.
Callers that need to mutate (build_context, package_context) already
deep_thaw before mutating, so the aliasing is safe.
The LE path still freezes and caches as before.
0 commit comments