John Ky

Switching to `haskell-language-server` is also a viable work-around. In this case, make sure you generate `hie.yaml` files if you have any issues.

Before: ``` Usage: cardano-cli transaction build-raw [--byron-era | --shelley-era | --allegra-era | --mary-era | --alonzo-era] (--tx-in TX-IN [--tx-in-script-file FILE [ (--tx-in-datum-file FILE | --tx-in-datum-value JSON VALUE) (--tx-in-redeemer-file FILE | --tx-in-redeemer-value...

Annotations allow me to output HTML to describe the structure of the document that is being rendered to better understand layout bugs:

The idea behind the new style is it prefers compact display such as this: ``` [--json-metadata-no-schema | --json-metadata-detailed-schema] ``` But that if the length of the group should exceed the...

I created another version of this PR without the tracing annotations so it's clearer where the substantive differences are. https://github.com/pcapriotti/optparse-applicative/pull/429

Mentioning https://github.com/pcapriotti/optparse-applicative/issues/273 and @quchen because the PR involves switching to `prettyprinter`.

I had a go yesterday, but it seems `ansi-wl-pprint` has different layout rules for its combinators and I wasn't successful. I pointed out tracing mainly to show that annotations are...

> What specifically do annotations buy us in case of optparse-applicative in production setting? What kind of richer console output are we looking for? I was thinking perhaps HTML help...

> With regards to the actual rendering changes: I'm quite sure we should rethink how we do the hang when the program name and subcommand names are long as squashing...

Another perhaps out-there idea would be to use rainbow parens 😁 .