Radioactive Butthole

It burns when I poop

  • 0 Posts
  • 99 Comments
Joined 21 days ago
cake
Cake day: November 28th, 2024

help-circle
  • That’s kind of my point though. For being made specifically for the purpose of being machine readable, its kind of a pain in the ass to work with.

    I want a command line utility where I can just

    xmlquery --query 'some/query' --file foo.xml --output foo-out.xml
    

    or in python

    
    import xml
    
    with open("foo.xml", "r") as file:
        data = xml.load(file.read())
    

    That’s the amount of effort I want to put into parsing a data storage format.


  • Meh. I just wish XML was easier to parse. I have to shuttle a lot of XML data back and forth. As far as I can tell, the only way to query the data is to download a whole engine to run a special query language, and that doesn’t really integrate into any of my workflows. JSON retains the hierarchy and is trivially parsed in almost any programming language. I bet a JSON file containing the exact same data would be much smaller also, since you don’t list each tag twice.


















  • The selling point is that it is immutable, not that it uses snaps (which it does). Fedora does the same thing with Silverblue and IoT. You don’t install rpms, you install flatpaks. You can install rpms, but you’re not really meant to.

    Since Canonical refuses to get onboard with flatpak (for now) they use snaps instead of debs, but snaps aren’t the direct appeal.

    The whole idea is that you have a core system in a known configuration. Updating the system just means using a different image. If an update fails, then you just roll back to the last good configuration. Bazzite uses this to nice effect too.

    There are a lot of advantages to end users and enterprise admins with systems in this configuration.