Zef interpreter after Optimization #17: Specialize toString
Before this change, doing x.toString would result in a slow path call that cascades through a bynch of Symbol* comparisons to realize that weβre using the toString operator.
This adds inline caching to Dot to handle the toString case.
Also, this changes exactly how we do int-to-string conversion so that we avoid unnecessary allocations.
2.7% speed-up.
This is displaying a diff from the previous version of the Zef interpreter.