ra’s avatarra’s Twitter Archive—№ 30,374

      1. hot off the benchmarks: CLR DynamicMethods are *not* significantly slower than static methods, ~3ns/call gist.github.com/nasser/0f63f6ea62a6c37d543fd9c1eb5bd1b1
        Permalink On twitter.com Twitter logo ra’s avatar ❤️ 3 Favorites Mood 0
    1. …in reply to @ra
      this is exciting because emitting DynamicMethods could be more flexible than emitting types at runtime
      Permalink On twitter.com Twitter logo ra’s avatar ❤️ 1 Favorite Mood +3 🙂
  1. …in reply to @ra
    a potential compilation strategy: DynamicMethods at run/dev time, static types/methods when “built”. MAGIC makes it easy to swap.
    On twitter.com Twitter logo ra’s avatar Mood +1 🙂
    1. …in reply to @ra
      many optimizations are defeated by supporting live coding, i.e. “what if the function changes later?” which is kind of useless when exported
      Permalink On twitter.com Twitter logo ra’s avatar ❤️ 1 Favorite Mood -1 🙁