A picture of a car with a license plate that reads LUAAAAA.

  • Melusine@tarte.nuage-libre.fr
    link
    fedilink
    Français
    arrow-up
    7
    ·
    1 day ago

    My main pain point with python at work. I lpke to write shitty scripts with it, but django was a mistake (the ORM opens then close connection to db with each request)

      • Melusine@tarte.nuage-libre.fr
        link
        fedilink
        Français
        arrow-up
        2
        ·
        20 hours ago

        I do, and still a pain in the ass because mypy is an afterthought of python (if your models aren’t in a migration then it’s an error)

          • Melusine@tarte.nuage-libre.fr
            link
            fedilink
            Français
            arrow-up
            1
            ·
            20 hours ago

            All our codebase is django and the admin is nice, and my manager is more conservative on the stack choice for the company XD

            • lime!@feddit.nu
              link
              fedilink
              arrow-up
              1
              ·
              19 hours ago

              yeah that’s fair. and migrating a codebase to being typed is a nightmare. it’s worth it, but fixing all the little edge cases that you didn’t know you had, or just boxing with the type checker when it freaks out, is an extremely frustrating experience.

    • marcos@lemmy.world
      link
      fedilink
      arrow-up
      3
      ·
      edit-2
      1 day ago

      You can set a connection pool. Or use a database proxy.

      EDIT: Oh, you are using Django for shell scripts?

      • Melusine@tarte.nuage-libre.fr
        link
        fedilink
        Français
        arrow-up
        1
        ·
        20 hours ago

        Yeah, but the python runtime still needs to open sockets. We have some kafka listeners using the ORM and we got the bad surprise of openning too many connection because django would keep opening them without closing them

        We used it for some batch processing and to have the admin for some managing (specially for retries)