1. 43
  1. 5

    I yearn for the day when Go removes all mentions of sort.Interface lol. Thanks for work BTW.

    1. 3

      I actually always really appreciated the cleverness of sort.Interface. It really showed what the language designers were going for.

      1. 4

        What were they going for?

        1. 2

          They were, as far as I can tell, trying to show the power of nominative types combined with interfaces.

    2. 3

      They finally fixed the ballast thing!

      1. 2

        I plan to continue to allocate an extra 100MB slab, for luck.

      2. 1

        What does that sort.Find do that is new ? Does it just give index number between 0 and N and generate those index by binary searching ? That’s IMO really weird interface ? Why not take a slice and a comparison closure taking one or two Ts and keeping the way the indexing works private ?

          1. 1

            Sure, those are the APIs I would expect. So I don’t understand what the sort.Find does that is “new” ? (except presenting a definitely weird interface to its user)

            It’s the “new” part of the statement I don’t understand. (and to a lesser degree, the interface design)

            1. 1

              It’s not “new” in that sense. Rob Pike just felt like sort.Search was too awkward, so he proposed a new function that’s maybe 10% less awkward. It’s the same functionality either way.

          2. 2

            You can use it for things other than slices. For example, you could use it to implement something like git-bisect