fix: defer const normalize in coherence mode by cordx56 · Pull Request #158205 · rust-lang/rust

cordx56 · 2026-06-21T08:18:14Z

I'm not confident that this fix is the best and desired in the long term point of view, but I'm confident that this does not do anything unsound.

Fixes #157937, by deferring normalization of associated const, except for anon, in coherence mode.
This fix prevents reporting impls, including associated const, are not overlapping.

rustbot · 2026-06-21T08:18:17Z

changes to the core type system

cc @lcnr

rustbot · 2026-06-21T08:18:19Z

Thanks for the pull request, and welcome! The Rust Project is excited to review your changes, and you should hear from @petrochenkov (or someone else) some time within the next two weeks.

Please see the contribution instructions for more information. Namely, in order to ensure the minimum review times lag, PR authors and assigned reviewers should ensure that the review label (S-waiting-on-review and S-waiting-on-author) stays updated, invoking these commands when appropriate:

@rustbot author: the review is finished, PR author should check the comments and take action accordingly
@rustbot review: the author is ready for a review, this PR will be queued again in the reviewer's queue

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 73 candidates
Random selection from 18 candidates

lcnr · 2026-06-21T22:12:29Z

r? lcnr

cordx56 · 2026-07-08T08:38:57Z

@rustbot review

lcnr

sorry for the late review 🙇

avoided this PR a bit as thinking about gce is annoying

View changes since this review

lcnr · 2026-07-15T08:58:27Z

                    // ripped out soon so this shouldn't matter soon.
-                    StructurallyRelateAliases::No if !tcx.features().generic_const_exprs() => {
+                    StructurallyRelateAliases::No
+                        if !tcx.features().generic_const_exprs() || defer_in_coherence =>


hmm, can you change this to instead be self.infcx.next_trait_solver() || !tcx.features().generic_const_exprs()

We don't support gce with the new solver but currently have the new solver in coherence if gce is enabled, so this should have the same results as your PR while being easier (for me) to reason about

I see, and thank you for reviewing!

I'm running tests and will push a revised version!

Wait, the ui test failed if I replaced my change with self.infcx.next_trait_solver() || !tcx.features().generic_const_exprs().

The case is tests/ui/const-generics/issues/issue-89304.rs that uses anon const kind like:

impl<const T: usize> From<GenericStruct<T>> for GenericStruct<{T + 1}> { ... } impl<const T: usize> From<GenericStruct<{T + 1}>> for GenericStruct<T> { ... }

that test should fail 🤣

From<W<A>> for W<{ expr involving A }> and From<W<{ expr involving A }>> for W<A> may overlap unless we do reasoning about the generic expression.

The compiler currently does not reason about the fact that T + 1 is larger than T, so that (T, T + 1) and (U + 1, U) can never be equal.

As in, while this example could be accepted in theory, the reason we do so right now is due to a bug and not because we correctly reason about generic expressions

Okay, so... //@ check-pass at line 1 on tests/ui/const-generics/issues/issue-89304.rs is actually caused by a bug and currently the compiler does not check this overlap accurately, right?

How should we handle the case like this? I don't know the process to decide these regressions in the Rust dev.
Should this be discussed in some other place? I've joined rust-lang Zulip.

exactly

for GCE it'd be fine to just delete the test I think

more generally, often it's one or a combination of

reopen the original issue linked from the test

open a new issue if the failure of the test changed

mark the test as //@ known-bug

update the comment in the test file to the new (intended) behavior

cry

generic_const_exprs will be replaced by generic_const_args soon. We'd otherwise update the comment and expected outcome of this test and keep it around, but I think here deleting is easier 🤷

I see, so I will delete the test, commit it, and push it. Thanks for reviewing!

…ted (GCE)

rustbot · 2026-07-20T15:21:12Z

This PR was rebased onto a different main 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.

cordx56 · 2026-07-20T22:05:25Z

@rustbot review

I changed the code to resolve conflict.

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jun 21, 2026

rustbot assigned petrochenkov Jun 21, 2026

rustbot assigned lcnr and unassigned petrochenkov Jun 21, 2026

This comment has been minimized.

Sign in to view

lcnr added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 6, 2026

cordx56 force-pushed the fix/const-coherence branch from 2054520 to 134d2ce Compare July 8, 2026 08:30

This comment has been minimized.

Sign in to view

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 8, 2026

lcnr reviewed Jul 15, 2026

View reviewed changes

rustbot added A-tidy Area: The tidy tool T-bootstrap Relevant to the bootstrap subteam: Rust's build system (x.py and src/bootstrap) labels Jul 20, 2026

This comment has been minimized.

Sign in to view

cordx56 added 4 commits July 21, 2026 00:15

fix: defer const normalize in coherence mode

67252fe

fix(test): remove the test that should not be passed and will be dele…

32a1b51

…ted (GCE)

fix: check whether next_trait_solver which does not support GCE

b882493

chore(test): remove problematic test from issues.txt in tidy

7e186a8

cordx56 force-pushed the fix/const-coherence branch from 20b9a4e to 7e186a8 Compare July 20, 2026 15:21

Uh oh!

Uh oh!

Conversation

cordx56 commented Jun 21, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jun 21, 2026

Uh oh!

rustbot commented Jun 21, 2026

Uh oh!

lcnr commented Jun 21, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

cordx56 commented Jul 8, 2026

Uh oh!

lcnr left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

rustbot commented Jul 20, 2026

Uh oh!

cordx56 commented Jul 20, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

cordx56 commented Jun 21, 2026 •

edited by rustbot

Loading

lcnr left a comment •

edited by rustbot

Loading