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: memoize collectSubfields
Collecting subfields occurs after resolving a field's value and before resolving subfield values. This step collects fragment spreads and checks inline fragment conditions. When fetching a list of things, this step is computed with the same inputs and expecting the same outputs for each item in the list. Memoizing ensures the work is done at most once per type returned from a list.
I tested this against the introspection query (which is both synchronous and complex) against a large schema and saw a ~15%-25% reduction in runtime. In practice I don't expect most queries to see this level of speedup as most queries are limited by backend communication and not execution overhead.
* Include context in memoization, Factor out memoize3
0 commit comments