don't touch closure_substs.parent_substs() in borrowck

[aliemjay]

... They bite!

An alternative implementation of #98835 that doesn't add any additional constraints on closure_substs.parent_substs in order to not change the behavior of opaque type inference. See #100372.

See the comment on the test for more details.

Fixes#100267.

r? types

[spastorino]

r? types

[bors]

☔ The latest upstream changes (presumably #98559) made this pull request unmergeable. Please resolve the merge conflicts.

[jackh726]

Now that ReEmpty has removed, does this PR work?

[PureWhiteWu]

Hi, is there any progress on this bugfix?
This bug has caused us unable to use the latest nightly version for one and a half month.
Thank you all for your work!

[apiraino]

I think this is waiting on @aliemjay to check this comment. Please request another review with @rustbot ready, thanks!

@rustbot author

[jackh726]

I went ahead and rebased this. It's not clear we should land this at this point. The only behavior difference I see is to the src/test/ui/type-alias-impl-trait/closure_wf_outlives.rs test, where we no longer require bounds on the GAT parameters (except for T: 'static).

[jackh726]

Let me go ahead and nominate.

[lcnr]

https://rust-lang.zulipchat.com/#narrow/stream/326132-t-types.2Fmeetings/topic/2023-01-18.20meeting/near/322078253

unnominating with the expectation that the PR is now unnecessary as the added tests already seem to pass. If not feel free to update the PR description and renominate 🤔

[jackh726]

@oli-obk can you look at src/test/ui/type-alias-impl-trait/closure_wf_outlives.rs and let me you think this change here is good or not?

[oli-obk]

looks good, but the now-passing parts should be pulled out into a separate check-pass test

[Dylan-DPC]

@aliemjay any updates on this?

[JohnCSimon]

@aliemjay

Ping from triage: I'm closing this due to inactivity, Please reopen when you are ready to continue with this.
Note: if you are going to continue please open the PR BEFORE you push to it, else you won't be able to reopen - this is a quirk of github.
Thanks for your contribution.

@rustbot label: +S-inactive