From hackernews π
One thing that keeps surprising me about Babashka is just how fast it is. At Nubank we have this giant collection of bash scripts called "nucli" that cover all sorts of development-time and operational tasks. I've been spending a few days trying to make it so that we can use http://github.com/hlship/cli-tools to define those commands in Babashka instead. I'm not about to rewrite the 1700+ scripts we already have, so at startup, my Babashka master-script reads the directory structure to find those legacy scripts and then reads each one to find documentation describing the script's purpose and usage. 2400 directories containing 9800 files. Startup time on my M1 Mac: 0.3 seconds, except I cache the results of the search so later startups are faster (.16 seconds). I have years of bias against scripting in Java and Clojure due to slow startup times. But in Babashka, if you are merely doing thousands of operations, execution time just doesn't seem to be a concern.
Well, my little experiment has been merged and now 1000s of Nubank developers may get a little taste of Babashka.
So ... what's the most elaborate thing you've written in Babashka? And how happy are you with the result? I'm sure @borkdude would love to hear.
β’ recently #clerk became bb-compatible which is quite a big and complex clojure project :-D it includes a liter version of tools.analyzer as well β’ recently aws-api also became bb-compatible
@hlship maybe send your reply to the channel as well to get more visibility, would love to get some replies here
already sent this to borkdude, and itβs not quite babashka, but https://codeberg.org/jyn514/flower uses SCI internally