• msage@programming.dev
      link
      fedilink
      arrow-up
      3
      ·
      1 year ago

      I need to buy this book!

      Though since it’s over a decade old, it will most likely be only a decoration.

      Then again, most mechanics didn’t change much.

        • msage@programming.dev
          link
          fedilink
          arrow-up
          1
          ·
          1 year ago

          If it describes low level mechanics, it’s still as relevant as ever.

          GUI is nice, but the game is still too complex for any user interface, and understanding it visually may not always be easy.

    • Kerb@discuss.tchncs.de
      link
      fedilink
      arrow-up
      2
      ·
      1 year ago

      Aw man, its a custom design.

      I’d love the classic animal template but with a kea holding a steppladder.

  • Eylrid@beehaw.org
    link
    fedilink
    arrow-up
    5
    ·
    1 year ago

    Finally an up to date edition! I’m still using the old book, Copying and Pasting from StackOverflow

  • nibblebit@programming.dev
    link
    fedilink
    arrow-up
    4
    ·
    1 year ago

    Hey! Why don’t you try to build a basic github action from scratch and deploy it 40 times before it’s green, waiting 10 minutes between builds to discover how each task creatively interprets directory path notation!

    • baum@programming.dev
      link
      fedilink
      arrow-up
      3
      ·
      1 year ago

      This is so true, I use copilot for things like actions and dockerfiles so I don’t feel an urge to bash my brains out more than I use it for code gen

    • ruffsl@programming.dev
      link
      fedilink
      arrow-up
      1
      ·
      1 year ago

      Get out of my head! Out I say!

      You’d think there’d be a better way of testing CI configs locally, like using selfhosted runners, but no, they need me to push changes to origin so that GitHub can then trigger the job on my local docker host. At least CircleCI gave us CLI to validate and partly run jobs from config files locally, … but only partly.

  • Sleeping@programming.dev
    link
    fedilink
    arrow-up
    3
    ·
    1 year ago

    I wonder if someone has already created a bot that asks chat-GPT to create a piece of code, grabs it to try and run it, and then just goes back and forth with chat-GPT to fix the errors. Now the code would probably be a complete mess, but I wonder if non-coders could use it to create helpful one-off tools they can’t find anywhere else.

    • Deebster@lemmyrs.org
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 year ago

      It definitely exists - there’s something that goes:

      1. generates code
      2. feed code back in and ask ChatGPT if it has any bugs (yes/no answer)
      3. if no, done, otherwise feed code back in and ask ChatGPT to fix it
      4. goto step 2

      I can’t find the link now, but you might also be interested in Auto-GPT.

  • meli nasa@feddit.de
    link
    fedilink
    arrow-up
    3
    ·
    1 year ago

    New: AI-powered unit testing framework that automatically corrects your code so that all tests pass! Perfect to integrate into your CI pipeline to run before deploy!