Here's my take on a reproducible benchmark for property accesses. What do others think about this? I picked a few representative properties (mostly arbitrarily).
My hope is that we can then implement changes to the property access code and make reasonable decisions regarding perf regressions.
Here's my take on a reproducible benchmark for property accesses. What do others think about this? I picked a few representative properties (mostly arbitrarily).
My hope is that we can then implement changes to the property access code and make reasonable decisions regarding perf regressions.