haskell-awk VS pretty-simple

I use a smooth awk-like tool called hawk for similar reasons and it sure is nice, can recommend.

Looking for a few projects that make use of it, I found accelerate, hawk, polysemy and pretty-simple, so I'll be interested to poke around in their code and see how they have things set up.

For easy, rapid development. If I'm processing text at the command-line and I realize that I need a Unix command-line tool which doesn't yet exist, say, one which reverses each line, I can very easily pipe my text into ghc -e 'interact (unlines . reverse . lines)' (or even more easily as hawk -md L.reverse) and move on to the next step of my text manipulation. If I had to open a text editor, create a new Haskell project, write the code which grabs the input, processes the lines, and then outputs the result, then this amount of friction is large enough that it's not worth it for a one time task. I'd probably find a different way to do it which is less expedient than the interact solution but more expedient than creating a new Haskell project.

hint maintainer here! To use dependencies, you need to use unsafeRunInterpreterWithArgs to pass extra -package-db arguments to ghc. In my hawk project, I'm trying hard to use the same package database in which hawk itself was installed, so I wrote a bunch of code to detect which installation method was used and to figure out the package database folder from there.

You can use https://hackage.haskell.org/package/pretty-simple for that. It's pretty good and it also adds color.

Looking for a few projects that make use of it, I found accelerate, hawk, polysemy and pretty-simple, so I'll be interested to poke around in their code and see how they have things set up.