fix: offer on path_type for add_lifetime_to_type

[A4-Tacks]

Example

struct Foo {
    a: &'_ i32,
    b: Foo<'_$0>,
}

Before this PR

Assist not applicable

After this PR

struct Foo<'a> {
    a: &'a i32,
    b: Foo<'a>,
}

[ChayimFriedman2]

I think it's better to check either for ast::Lifetime or for a without-lifetime ast::RefType. Technically it will be nice if the assist will also trigger for missing lifetime params in ADTs, but I think this is better done by having a diagnostic for this with a quickfix (that can supersede this assist).

View changes since the review

[A4-Tacks]

I think it's better to check either for ast::Lifetime or for a without-lifetime ast::RefType.

The text range of '_ is too small, which may not be conducive to user use

but I think this is better done by having a diagnostic for this with a quickfix (that can supersede this assist).

After the diagnostics stabilize, consider migration

Technically it will be nice if the assist will also trigger for missing lifetime params in ADTs

Good idea

[ChayimFriedman2]

The text range of '_ is too small, which may not be conducive to user use

IMO it's confusing if we offer this on unrelated parts of the path.

[A4-Tacks]

This assist always needs to be applied or not triggered because it is unlikely to have a false positive

Perhaps it's a good idea for the entire Adt application, what do you think?

[ChayimFriedman2]

This assist always needs to be applied or not triggered because it is unlikely to have a false positive

What do you mean?

[A4-Tacks]

For an Adt in development, exists a missing lifetime argument:

• When exists a lifetime parameter: assist not applicable
• When non-exists a lifetime parameter: you always need to add lifetime parameters and arguments
  • uses this assist: At most, it's just an additional renaming
  • not uses this assist:
    • add a lifetime parameter: assist not applicable because exists a lifetime parameter
    • add a lifetime arguments: assist not applicable because not-exists missing lifetime

[A4-Tacks]

Anyway, let's implement it this way first. I have placed an XXX comment

[rustbot]

This PR was rebased onto a different master 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.