The work on FLUID-5249 very substantially improved the dispatch of invokers, by perhaps a factor of 60, bringing the overhead to perhaps 3-4 microseconds on typical desktop machines. This allowed us to eliminate the so-called "fast invokers" (FLUID-4922) from configuration since all invokers became equivalently and maximally fast. In certain cases (where IoC references are sourced from either {that} or {arguments}) the dispatch is allocation-free.

We need to apply the same optimisations in fluid.dispatchListener - to i) apply the two-stage options expansion and ii) guard any logging statements using fluid.defeatLogging