Should shell history files be STATE, CACHE, or DATA? about xdg-ninja HOT 3 CLOSED

Personally, I think shell history is important enough to be kept in the DATA dir.

There's several projects with thousands of stars that save more info about your history, adds an easier search mechanism, syncs it between machines, etc., so I think I'm not alone on this.

I'd say we stick with it for zsh, and change it to DATA for bash. Let me know your opinion, since I don't want to base this on just on my own opinion.

It's a tough call. XDG_STATE_HOME was only recently introduced, with the first proposal being in December of 2016 and implementation into the specification in May of 2021.

I believe the rule of thumb is that state is machine-specific, while data is cross-machine. Since shell history is most commonly used to build completion rules/suggestions, you wouldn't want it transferred to other machines with a different context, or worse, leak sensitive data.

I am going to propose standardizing on STATE, since it appears to be the current specifications standard for history files, and seems to follow with the initial proposal.

Looking forward to any other thoughts or opinions before settling one way or another. :)

from xdg-ninja.

Personally, I think shell history is important enough to be kept in the DATA dir.

There's several projects with thousands of stars that save more info about your history, adds an easier search mechanism, syncs it between machines, etc., so I think I'm not alone on this.

I'd say we stick with it for zsh, and change it to DATA for bash. Let me know your opinion, since I don't want to base this on just on my own opinion.

from xdg-ninja.

Convinced me. If someone cares enough about their history files, they probably know what to do anyway.

Feel free to submit a PR to change this for bash and zsh, otherwise I'll do it myself tomorrow.

from xdg-ninja.