• poVoq@slrpnk.net
    link
    fedilink
    arrow-up
    5
    ·
    2 years ago

    The massive speed improvements are super exciting. Could be a php 7.x like moment for Python.

      • poVoq@slrpnk.net
        link
        fedilink
        arrow-up
        2
        ·
        edit-2
        2 years ago

        Its not that bad right now, and 10-60% speed improvements makes it at least on par with other interpreted languages again. You can’t compare it to compiled languages of course, but it is not always a good idea to use those.

        • holdengreen@lemmygrad.ml
          link
          fedilink
          arrow-up
          1
          ·
          2 years ago

          Not if you compare it to languages like Lua, Javascript, LISP. Those languages can get really close to native for general code.

          Not suggesting Python needs to become more like those languages, which generally don’t have the same nice features Python does.

          But I’m saying it should maybe be a goal to get much closer to them without having to write a C module.

  • pingveno
    link
    fedilink
    arrow-up
    2
    ·
    2 years ago

    That dataclass_transform looks pretty cool. There has traditionally been some shortcomings between Python’s type annotations and more complex behavior of certain classes. For example, Django has complicated behavior in its ORM governed by metaclasses. Currently that requires a custom bespoke plugin for each type checker. dataclass_transform will instead allow them to write the one chunk of code that will work everywhere.