Do not suggest `unwrap` on `Some(T)` where `T` is expected

[oli-obk]

Currently

let i: i32 = Some(42);

reports

2 |     let i: i32 = Some(42);
  |                  ^^^^^^^^ expected i32, found enum `std::option::Option`
  |
  = note: expected type `i32`
             found type `std::option::Option<{integer}>`
  = help: here are some functions which might fulfill your needs:
          - .unwrap()
          - .unwrap_or_default()

I think Option and Result are special enough to warrant customized error messages. In this case simply suggesting to remove the Some

[nikomatsakis]

See #42929 -- I think that we should not be giving such broad suggestions based solely on types. It's too hard to tell the difference of what is appropriate or inappropriate.

[zackmdavis]

It's worth stressing that while this makes sense for the Some literal in the given example, in more complicated situations, it's entirely plausible for .unwrap() to be the right suggestion when a variable of type Option<T> is given where T is expected. (Discussion on #42929 seemed to blur the distinction.)

[estebank]

Whitelist implemented in #46461:

error[E0308]: mismatched types
 --> src/main.rs:2:14
  |
2 | let i: i32 = Some(42);
  |              ^^^^^^^^ expected i32, found enum `std::option::Option`
  |
  = note: expected type `i32`
             found type `std::option::Option<{integer}>`

error: aborting due to previous error