Interesting, does this demonstrate better perf than some of the other approach? Last time I checked IL emit was the fastest but had an extreme first time cost. Where does this approach fall in the startup perf vs. per-invocation perf?
Happy to merge it either way to demonstrate a variety of approaches.
Interesting, does this demonstrate better perf than some of the other approach? Last time I checked IL emit was the fastest but had an extreme first time cost. Where does this approach fall in the startup perf vs. per-invocation perf?
Happy to merge it either way to demonstrate a variety of approaches.