bash-cache VS bkt

This is usually what a caller expects, but it's definitely surprising when it isn't. I encountered this with my bash-cache library because simpler approaches to caching commands will use command substitutions which turns out to be lossy (because it drops any trailing newlines).

Is the goal of your cache function to cache a command if it's been run before? If so you might like to use my bash-cache library :) Even if it's not quite what you're looking for you can poke around the code for some inspiration, it implements a function decorator pattern and demonstrates some hairy uses of eval.

For another pure-bash solution (though it writes to disk, not in-memory arrays) check out my bash-cache library :) it's an older project that motivated bkt but I still use both. There's some terrifying bash patterns in bash-cache as well!

I'm pretty proud of my bash-cache utility, which implements a decorator pattern in Bash. Define a function, pass it to bc::cache, and like magic the function is wrapped with a caching layer.

Hi, author of bkt here. I use fzf all the time and combine it with bkt to make --preview faster. I hope you find this pattern useful!

Alternatively, check out bkt (also mine), which is like bash-cache but for subprocesses instead of shell functions.

Instead, using bkt ensures each request is only sent once and all subsequent calls return locally cached results:

Enjoy! If you find it useful please let me know how you're using it!

Honestly, any reasonably complex bash script uses (or should be using!) arrays, so just poke around and you'll probably see them. Here's a fairly straightforward script of mine with several arrays: https://github.com/dimo414/bkt/blob/master/benchmark.sh

Glad to hear it! I hope you find it helpful :) If you have a few minutes please share how you're using bkt, it'll help me and others!