xdg-ninja VS boxxy

As a user, these variables make my experience simpler. I'm not going to argue that these specifications should be followed by all, because I know there are many users who are committed to dying on the hill that is their cluttered home directory. However, the existence of these variables is not a deterrent to users who do not want to use the specification, as many applications will want to use your home directory anyway. If the existence of these variables made the specification strictly followed, projects like xdg-ninja (https://github.com/b3nj5m1n/xdg-ninja) would have no reason to exist.

Regardagin cargo (and other tools), I've had some success with following suggestions from https://github.com/b3nj5m1n/xdg-ninja

It's not as bad as it used to be. And ther's even software that can help you with that: https://github.com/b3nj5m1n/xdg-ninja

Check out xdg-ninja

https://github.com/b3nj5m1n/xdg-ninja

This utility has been a lifesaver to clean up my home directory.

Stuff like [xdg-ninja](https://github.com/b3nj5m1n/xdg-ninja) helps but... at one point my home is still a mess.

For the ones that adhere to XDG specs, there are tools that can help transition to using these XDG directories and avoid breaking programs - like XDG Ninja - but it's still a sort of manual process that doesn't cover all dotfiles and applications.

There are shell scripts like xdg-ninja that can help with this:

https://github.com/b3nj5m1n/xdg-ninja

View on GitHub

I made a tool to control where Linux applications put their files via bind mounts and environment variables: https://github.com/queer/boxxy

I've heard that it's made it as far as university HPC clusters to help control iffy code written by students; I'm glad I managed to make that stuff a bit easier for the people operating them.

This is a tool I’ve been working on for the last ~6 months. It was born out of frustration trying to package boxxy[1] for different distros.

This was originally to be presented at RustConf 2023 in a few days, under the talk “Repackage the World!,” but unfortunately my health took a sudden turn for the worse and I’m no longer able to give the talk.

peckish IS NOT a replacement for distro-specific packaging tools, and its packages WILL NOT be compliant with every distribution’s standards for packages.

peckish IS:

- a way to make a quick-and-dirty packages for distributing a program

- a CLI/library for manipulating the contents of Linux packages

The core abstraction is an in-memory filesystem behind an async std::fs-like facade[2], allowing packages to be manipulated with more-intuitive random-access I/O instead of putting up with (nested) streaming archive formats. This facade also is used for ex. enabling easier archive manipulation[3], a library for copying between facade implementations[4], and more. This bets that most packages will never be larger than memory, which in my testing is a safe bet even with Docker images as inputs.

Sorry for adding more YAML to the world :P

[1] https://github.com/queer/boxxy

[2] https://github.com/queer/floppy-disk

[3] https://github.com/queer/flop

[4] https://github.com/queer/disk-drive

boxxy

I am not sure if you know boxxy: it is a handy tool that uses linux userspace namespaces to wrap requests for a file to a proper (configurable) location.

macOS + systemd? If on Linux this tool allows you to move config dir and file locations transparently. Pretty neat. https://github.com/queer/boxxy