resolve: Do not consider traits from ambiguous imports to be in scope

[petrochenkov]

One of unblocking steps for #145108.
Fixes #147992.

[rustbot]

r? @madsmtm

rustbot has assigned @madsmtm.
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

[petrochenkov]

@bors try

[petrochenkov]

cc @LorrensP-2158466

[rust-bors]

☀️ Try build successful (CI)
Build commit: ed60662 (ed606620234c6b50f2414162c4bbabd5f4a925f3, parent: f5e2df741b4a9820a7579f0c8eccc951706a8782)

[petrochenkov]

@craterbot check

[craterbot]

👌 Experiment pr-147995 created and queued.
🤖 Automatically detected try build ed60662
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🚧 Experiment pr-147995 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[madsmtm]

Nit: Could you add #[rustfmt::skip] to this and test2, I'd be worried about the test no longer testing the right thing if someone decided to reformat these.

[madsmtm]

This makes sense to me, and I agree that it needs a crater run.

View changes since this review

[craterbot]

🎉 Experiment pr-147995 is completed!
📊 135 regressed and 3 fixed (721923 total)
📊 1777 spurious results on the retry-regessed-list.txt, consider a retry¹ if this is a significant amount.
📰 Open the summary report.

⚠️ If you notice any spurious failure please add them to the denylist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

Footnotes

1. re-run the experiment with crates=https://crater-reports.s3.amazonaws.com/pr-147995/retry-regressed-list.txt ↩

[petrochenkov]

All the root regressions (31 of them) are

error[E0599]: no method named `method` found for type `Type` in the current scope

, as expected.
Also many non-root regressions in crates depending on various versions of sophia_api and polars-plan.

I guess we need to somehow preserve these traits in scope, but keep them poisoned, so an error (or a lint) can be reported later during type checking.

[madsmtm]

I guess we need to somehow preserve these traits in scope, but keep them poisoned, so an error (or a lint) can be reported later during type checking.

Or a FCW, I think it makes sense to break this someday.

[lcnr]

how difficult is the alternative of treating overlapping trait wildcard imports as use module::Trait as _; instead

View changes since this review

[petrochenkov]

I guess we need to somehow preserve these traits in scope, but keep them poisoned, so an error (or a lint) can be reported later during type checking.

Or a FCW, I think it makes sense to break this someday.

FCW will also have to be reported in type checking, can't do that without keeping the poisoned traits in scope.

[petrochenkov]

how difficult is the alternative of treating overlapping trait wildcard imports as use module::Trait as _; instead

View changes since this review

I'll try to implement both the alternatives and look what's simpler.

[petrochenkov]

Closing in favor of #149058.