Hey Andy, On Mon, Sep 17, 2018 at 6:18 PM Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote: > I think it’s fine for later. It’s potentially useful for benchmarking and debugging. I went ahead and added it anyway in the end. It was really quite easy to do, and it sets a good template for future primitives that are added to Zinc. Meanwhile, I've fully triaged the regression Martin found, and in the same test he performed, the Zinc code now runs faster than the old code (which should be no surprise). I've also finished getting rid of -include, per your suggestion, and modularizing each algorithm to be loadable independently, per your suggestion. All and all, v5 is turning out really well, and I'll probably submit it sometime soon. My tree was just added to the kbuild test bot, so I'll hopefully get an email about any breakage *before* I submit this time. Regards, Jason
