Suggest using deref in patterns

[uellenberg]

Fixes #132784

This changes the following code:

use std::sync::Arc;
fn main() {
    let mut x = Arc::new(Some(1));
    match x {
        Some(_) => {}
        None => {}
    }
}

to output

error[E0308]: mismatched types
  --> src/main.rs:5:9
   |
LL |     match x {
   |           - this expression has type `Arc<Option<{integer}>>`
...
LL |         Some(_) => {}
   |         ^^^^^^^ expected `Arc<Option<{integer}>>`, found `Option<_>`
   |
   = note: expected struct `Arc<Option<{integer}>>`
                found enum `Option<_>`
help: consider dereferencing to access the inner value using the Deref trait
   |
LL |     match *x {
   |           ~~

instead of

error[E0308]: mismatched types
 --> src/main.rs:5:9
  |
4 |     match x {
  |           - this expression has type `Arc<Option<{integer}>>`
5 |         Some(_) => {}
  |         ^^^^^^^ expected `Arc<Option<{integer}>>`, found `Option<_>`
  |
  = note: expected struct `Arc<Option<{integer}>>`
               found enum `Option<_>`

This makes it more obvious that a Deref is available, and gives a suggestion on how to use it in order to fix the issue at hand.

[rustbot]

r? @BoxyUwU

rustbot has assigned @BoxyUwU.
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

[compiler-errors]

I know it's preexisting, but this snippet logic is unnecessary. You should be able to just give a suggestion like "*" that is inserted before the match scrutinee with span.shrink_to_lo().

[compiler-errors]

It may also look better if you change this into a "verbose" suggestion with span_suggestion_verbose.

[uellenberg]

I've gone back on this, because I'm not sure how to override a span in a suggestion using multipart. Please let me know if there's a way to do so.

[uellenberg]

I haven't gone through the tests yet, because there are still some problems that I haven't figured out yet. The logic for adding references still has to be redone, but that will come with solving the other issues.

First, I'm still not happy about the suggestion given above. I may just add a check for the simple case of having a reference, since I think that covers the majority of times it'll appear.

A bigger issue is that, when rustfix tries to apply the suggestions, it does so for each part of the match arm. For example, tests/ui/issues/issue-57741-dereference-boxed-value/issue-57741.rs fixes to **x, even though each suggestion only says to use *x. I can add a check for the arm index, but that happens really far up in the call hierarchy. Is there a better way?

[uellenberg]

The issues I mentioned above have been resolved.

[uellenberg]

I can also expand the suggestion to specify which Deref impls are being used to get from type A to B.

[uellenberg]

r? rust-lang/diagnostics

[oli-obk]

Instead of using a snippet, you can track a prefix and an optional postfix suggestion. This way you can build the multipart suggestion right when you need it while being able to modify the prefix as you need to

[uellenberg]

I haven't been able to implement this without rustfix applying the suggestion twice. Box deref suggestion's MachineApplicability could be downgraded, but that seems worse.

Relevant zulip: https://rust-lang.zulipchat.com/#narrow/channel/182449-t-compiler.2Fhelp/topic/Rustfix.20applying.20multipart.20suggestion.20twice

[uellenberg]

@rustbot review

[oli-obk]

Thanks for the thorough investigation. I think we'll need to adjust rustfix to also consider suggestion spans as overlapping if they touch

@bors r+

[bors]

📌 Commit 831f454 has been approved by oli-obk

It is now in the queue for this repository.

[oli-obk]

Whoops, clicked review request instead of dismiss on mobile by accident