Awk support.
Using only gawk, which outperforms the original bash program.
Making the tied values sorted alphabetically makes it slower than first version, but still largely outperforms the original bash program on my machine (don't have docker installed).
That's strange, I cannot reproduce the same behavior on a linux box (hosted on digital ocean) and on my macbook pro. It turns out that the current awk implementation is 3 to 4 times slower than the original bash program on linux, but on my mac it's exactly the inverse scenario :smile_cat:
That is unusual, I wouldn't expect such a big gap between the implementations. Should I merge this or do you want to keep working on it?
I'll work on it, I've got an OCaml implementation pending also. thanks!
Looking forward to your OCaml imeplementation :)