Suggest Explicit Lifetime for Associated Type Bindings

[veera-sivarajan]

Fixes #122025

This PR suggests an explicit lifetime annotation for associated type bindings. Previously, this error suggested an incorrect HRTB.

Please let me know if there's a better way to figure is_associated_type_binding. Also, this suggestion uses Applicability::MaybeIncorrect as the suggestion can be incorrect when a trait's input lifetimes are also missing.

[rustbot]

r? @cjgillot

rustbot has assigned @cjgillot.
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

[cjgillot]

There is existing diagnostic code in rustc_resolve::late::diagnostics. We should adapt and reuse add_missing_lifetime_specifiers_label and suggest_introducing_lifetime.

[fmease]

CC #123713 (comment) (last paragraph). The “proper” way to fix #122025 is to stash the diagnostic in rustc_resolve::late and to try_steal_modify_and_emit_err it during HIR ty lowering where we have the necessary tools available to check if a lifetime param would be unconstrained (cc validate_late_bound_regions, collect_constrained_late_bound_regions). However that might not be feasible to implement, needs experimentation.

I've only glanced at the changes made in this PR but you seem to be taking the alternative approach that I've dubbed “heuristic in late” in my linked comment. That's perfectly valid, too. My linked comments might provide some more food for thought :)

[bors]

☔ The latest upstream changes (presumably #123968) made this pull request unmergeable. Please resolve the merge conflicts.

[Dylan-DPC]

@veera-sivarajan any updates on this? thanks

[veera-sivarajan]

Still working on it. Thanks for checking.