Consideration to include a Pretty API about rio

[akhra]

The primary goal of RIO is not to include batteries, but to collect and enshrine established best practices. There's a bar of non-controversiality that has to be passed for something to be included.

This is largely subjective, of course; but the basic question is: among people who have a working familiarity with all the available options, is there one that stands out as the most common pick by a wide margin? If so, it may have a place in RIO. If not, it's better to leave it out until/unless a clear winner emerges.

[roman]

Fair enough, I'm not aware of any clear champions in this area, I'll close this ticket.

Thanks.

[mgsloan]

I agree that it would be nice to have a standard pretty printer. I also agree that nothing has become standard enough practice that. It seems to me like it would be nice to have a package that integrates rio's logging with some prettyprint package, and perhaps adds a typeclass if not already provided by the prettyprint package.

Stack has a rather custom setup, using annotated-wl-pprint and adding its own logic for annotation with ANSI formatting and pretty print utilities:

• https://github.com/commercialhaskell/stack/blob/master/src/Text/PrettyPrint/Leijen/Extended.hs (the ansi annotations support)
• https://github.com/commercialhaskell/stack/blob/master/src/Stack/PrettyPrint.hs (utilities for combining logging with + pretty print functions)

Not entirely happy with that prettyprint setup, but it works. I don't have concrete examples in mind right now, but in the past it's been somewhat error prone to get good layout.

This stuff predates quchen's prettyprinter stuff. I have an old note in my stack todo list to switch over to that. I haven't looked deeply but I'm thinking it would allow removing some of this code.