You had asked in my prior pull request if this was actually a performance improvement outside of the micro-bench we provided, and for our use-case it is. We have a graph with multiple lines on it that have thousands of points. We re-draw those lines when users brush/scroll (the lines are dynamic, so can't just pan/zoom with clipping). This function stood out as the biggest CPU issue in performance tests, which is what led to this optimization. We npm linke'd our mod into our project and did notice a real-world performance increase - helped about 10% in total CPU time for the entire graph, and moved this function out of the top offenders performance wise.

Okay. I’m happy to try it and see if anyone complains (or rejoices).