Firefox on Debian stable is so old that websites yell at you to upgrade to a newer browser. And last time I tried installing Debian testing (or was it debian unstable?), the installer shat itself trying to make the bootloader. After I got it to boot, apt refused to work because of a missing symlink to busybox. Why on earth do they even need busybox if the base install already comes with full gnu coreutils? I remember Debian as the distro that Just Wroks™, when did it all go so wrong? Is anyone else here having similar issues, or am I doing something wrong?

  • EddyBot@discuss.tchncs.de
    link
    fedilink
    arrow-up
    9
    ·
    5 months ago

    the work amount of backporting fixes which ARE already fixed in newer versions is also insane

    thats one of the reason why Arch Linux sticks to stable upstream versions, backporting is just not feasable on smaller teams

    • gbin@lemmy.ca
      link
      fedilink
      arrow-up
      1
      arrow-down
      2
      ·
      5 months ago

      I have been an Arch user for years now and anytime I touch a debian based distro it is such a headache: weird patched packages that don’t compile anything past or present, insta dependency hell with PPAs, package names of 200 characters because apt doesn’t have a good way to represent metadata… It made me a strong believer that trying to fight the bit rot and stick to the old stuff is counterproductive: a consistent head based development with a good community fixing bugs super quickly results in less hours of work fighting the paleolithic era dependencies, safer (as security fixes are faster to get in, packages are foreign to hackers and constantly changing etc), easier to find documentation as you don’t need to dig into history to find which option existed or not, recent stuff is also easier to support for the developers of the various packages as it is fresh in their minds. Another point is to look at it from a tech debt lens: either you fix your stuff to work with current deps now or you just accumulate tech debt for the next engineer to fix in a way larger and combining a mountain of breakages in the future that of course IT and SREs will never want to do until the 15y old software is a disaster of security issues…