Add pretty_print methods to Result and Maybe

[paul]

Adds #pretty_print to Result and Maybe, so PP looks good.

Examples:

>> pp M::Success()
Success()

>> pp M::Success(User.first)
Success(
 #<User:0x00007f414e3584a0
  id: 1,
  email: [FILTERED],
  name: "Paul",
  staff: true,
  automated_actor: false,
  settings: {},
  auth_uid: nil,
  auth_profile: nil,
  created_at: "2025-05-13 21:34:12.784955000 +0000",
  updated_at: "2025-05-13 21:34:12.784955000 +0000",
  discarded_at: nil>)

>> pp result.to_monad
Failure(
 #<Dry::Validation::Result
 {now: 2025-05-16 14:52:40.354675084 UTC +00:00}
 errors={chat: ["is missing"], user: ["is missing"], user_message: ["is missing"]}>)

>> pp M::Some(User.first)
Some(
 #<User:0x00007f414e35a8e0
  id: 1,
  email: [FILTERED],
  name: "Paul Sadauskas",
  staff: true,
  automated_actor: false,
  settings: {},
  auth_uid: nil,
  auth_profile: nil,
  created_at: "2025-05-13 21:34:12.784955000 +0000",
  updated_at: "2025-05-13 21:34:12.784955000 +0000",
  discarded_at: nil>)

>> pp M::None()
None

Also, Pry and IRB leverage PP for their output, and even can color it, so this lets it color the monads, and the contents of the monads:

Before:

After:

[flash-gordon]

I'm positive about this, but can you make it an extension?

[paul]

I can, but I'm curious as to why. Here's my case for it not being an extension:

1. It provides a nicer out-of-the-box DX for newcomers, or people using then through another gem without realizing it. If its an extension, then people need to know that 1) it exists and 2) they have to explicitly enable it. Plus, that's made all that much harder because there's no documentation about extensions. I've been using Dry::Monads in projects for many years, and I didn't even realize it had extensions.
2. It doesn't "cost" anything. I see the point of extensions for optional functionality that might change behavior, or pull in expensive dependencies. But this implements 4 methods in 30 lines of code. The presence these methods being defined but never used costs a minuscule amount of memory and no processing time.

If you still want it as an extension, that's fine, its your project. I'm skeptical of the tradeoff involved, and urge you to reconsider, or at least explain your decision.

[flash-gordon]

they have to explicitly enable it

I don't have strong feelings if it's enabled by default, I just don't like when code is cluttered with significantly different things. My suggestion is trying to extract representation-related stuff to a separate file, similarly to how inspection is handled in dry-types. That said, it's not obvious to me which approach is better in the long run, I'll play around with extension to see if I actually like it. I'll send a PR later today

[timriley]

Thanks for putting this together, @paul! And thanks for taking care of it, @flash-gordon :)

FWIW, I'd love to have the pretty printing support enabled by default. It's a great example of how we can make our gems better to interact with, and one I'd love to carry over to the rest of our gems too. ❤️

[flash-gordon]

@paul @timriley yeah, thanks for bringing it up. I spent some time fruitlessly trying to bend pp to print

Some[
  1,
  2,
  ... # many more values
  100
]

Instead of its defaults:

Some(
 [1,
  2,
  # ...
  100])

Alas, it turned out to be less trivial so I gave up.

I've finished the implementation for all monads in an auto-loaded extensions #190 I think it's good to go. If you folks don't have objections, I'll releases it later today