Improve suggestions when multiples tuples implement the same trait

[rperier]

View all comments

Fixes #152903

[rustbot]

r? @oli-obk

rustbot has assigned @oli-obk.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler, types
• compiler, types expanded to 74 candidates
• Random selection from 16 candidates

[rperier]

Perhaps a ui test dedicated to this use case, specifically might be helpful, imho. It was convered by any tests.

[mejrs]

Idea: can we style the type like rustdoc does: https://doc.rust-lang.org/nightly/std/fmt/trait.Debug.html#impl-Debug-for-(T,) , like this

(T₁, T₂, …, Tₙ)

I'm a bit worried that it's going to look like shit in a terminal, though...

[rperier]

Idea: can we style the type like rustdoc does: https://doc.rust-lang.org/nightly/std/fmt/trait.Debug.html#impl-Debug-for-(T,) , like this

(T₁, T₂, …, Tₙ)

I'm a bit worried that it's going to look like shit in a terminal, though...

I like this idea, the output would be more concise and simpler. Plus, I would not have to count the max number of tuples args, so the code would be simpler.

However, I agree with you, how is it about to be rendered in a terminal ? I am going to investigate

[rperier]

Idea: can we style the type like rustdoc does: https://doc.rust-lang.org/nightly/std/fmt/trait.Debug.html#impl-Debug-for-(T,) , like this

(T₁, T₂, …, Tₙ)

I'm a bit worried that it's going to look like shit in a terminal, though...

I like this idea, the output would be more concise and simpler. Plus, I would not have to count the max number of tuples args, so the code would be simpler.

However, I agree with you, how is it about to be rendered in a terminal ? I am going to investigate

Simply by using UTF8 ? UTF8 can handle these kind of notation, normally (exponent, indices, etc...)

[rperier]

(T₁, T₂, …, Tₙ)

https://play.rust-lang.org/?version=stable&mode=debug&edition=2024&gist=b5b574df5b35a52a002eca066d9b4ba4 ;)

[rperier]

What about something like this ? The rendering in a terminal is quiet good.
I have left n as indice of T but I could also display 12 , which is more correct.

[mejrs]

I have left n as indice of T but I could also display 12 , which is more correct.

I do think it would be nice to say "up to tuples of length 12" or something.

[rperier]

I have left n as indice of T but I could also display 12 , which is more correct.

I do think it would be nice to say "up to tuples of length 12" or something.

Like :

error[E0277]: `Foo` doesn't implement `Debug`
 --> tests/ui/suggestions/multiple_tuples_implement_same_trait.rs:8:17
  |
8 |     testing((1, Foo));
  |     -------     ^^^ the trait `Debug` is not implemented for `Foo`
  |     |
  |     required by a bound introduced by this call
  |
  = help: the following other types implement trait `Debug`:  
            ()
            (T₁, T₂, …, Tₙ) // up to tuples of length 12
            // Other relevant impls here
          and 5 others
  = note: required for `({integer}, Foo)` to implement `Debug`

?

[rperier]

@rustbot ready

[mejrs]

r? me

Thanks, great work. I love this message specifically being more concise.

How does this diagnostic compose with #[splat]? Can you add a test documenting what that looks like? If it's weird or doesn't work then no need to fix that necessarily, feel free to punt it towards the splat people :)

Also cc @estebank I know you care deeply about how errors show up in the terminal, is use of subscript acceptable to you?

View changes since this review

[rperier]

r? me

Thanks, great work. I love this message specifically being more concise.

How does this diagnostic compose with #[splat]? Can you add a test documenting what that looks like? If it's weird or doesn't work then no need to fix that necessarily, feel free to punt it towards the splat people :)

Also cc @estebank I know you care deeply about how errors show up in the terminal, is use of subscript acceptable to you?

View changes since this review

Sorry, I don't know what #[splat] is . It's a community crate or as part of the compiler ?

[mejrs]

Sorry, I don't know what #[splat] is . It's a community crate or as part of the compiler ?

It's an experiment, see https://rust-lang.github.io/rust-project-goals/2026/overloading-for-ffi.html and #153629

[rperier]

There are two branches actually, the case when all types implements the same traits , with less than 5 candidates and the case where there are more than 5 candidates. I did not changes anything to the case with less than 5 candidates, I have added a "concise output" for continuous arities (ranges from 0 to N or from K to N) where there are more than 5 candidates.

@rustbot ready

EDIT: in order to be sure that all tuples and continuous by step of 1, I had to sort the candidates. That's why I have updated the output of some tests, only the ordering output of these test have changed.

[mejrs]

It looks good to me, just a few nits

View changes since this review

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[rperier]

@rustbot ready

[mejrs]

@bors r+ rollup

View changes since this review

[rust-bors]

📌 Commit 507283d has been approved by mejrs

It is now in the queue for this repository.