Closed shajra-cs closed 7 years ago
I'd prefer not to lose the generality of the pretty-printing combinators, because that might make it harder to port future changes from the Haskell library, or to make any other changes that are desirable.
I'd be open to merging in a minimal pretty-printing library containing just the combinators we do use, if it was a general enough foundation to expand on if necessary, without needing to unpick much specialised code. Performance is of no great concern to me.
👍 yeah it also pulls in a bunch of other dependencies. I was just trying to pull this lib up to 2.12 and it's blocked on Kiama, which may in turn be blocked on other stuff. If they don't get a 2.12 release out shortly I may look into slicing it off.
Has anyone started this work? I might give a naive port of the Wadler paper a try if no one is close to a solution.
I wonder if scala.text
would get us what we need? It's deprecated but still in 2.12 so we're ok for a while. Worst case could copy-paste the source when it goes away for good.
@tpolecat It really depends how set on the exact same spacing @bmjames is. I've almost completed a simple version of what Kiama is based on. We'll see if that works and what @bmjames thinks of it. :laughing:
This was closed by #13. @bmjames Can we get a release with the Kiama dependency removed?
Sure, I have been intending to make one but I haven't got round to it yet. I'll close this issue once the release is done.
My current company for very social (not rigorously technical/legal) reasons is uncomfortable with the LGPL license pulled in by Kiama.
It doesn't seem like Kiama is used too much by this project, and something hand-rolled might be possible -- maybe at the loss of some performance or generality -- but this is just a command-line parsing tool, so hopefully it's not that big a deal.
If someone else (me) put in the work, would you be open to getting off of Kiama?