implement const Iterator for Range by Randl · Pull Request #156216 · rust-lang/rust

Randl · 2026-05-06T01:21:52Z

This allows the use of for i in i..n in const.

This I believe is interesting enough to be justified under #155816

Randl · 2026-05-06T01:29:06Z

Another issue I noticed is that while is const-stable, but for isn't. So in some places, reverting the const-hack would make currently const-stable functions const-unstable. Alternatively, we wait until Iterator is const-stable.

theemathas · 2026-05-06T02:37:04Z

Does this run into #148200 for types that implement non-const TrustedStep but implements const Step?

Randl · 2026-05-06T04:40:57Z

@theemathas
Can you give maybe an example you had in mind? I'm not sure I can think of one. If I understand correctly, there won't be any non-const functions involved in this case, since the entire implementation will be in the const Step?

theemathas · 2026-05-06T07:42:43Z

Maybe there's actually not a problem with specialization here? I suppose it's kind of fine to call a function that requires const TrustedStep even though the type only actually implements TrustedStep + const Step, since TrustedStep itself doesn't have any methods.

oli-obk · 2026-05-06T08:47:53Z

 #[stable(feature = "rust1", since = "1.0.0")]
-impl<A: Step> DoubleEndedIterator for ops::Range<A> {
+#[rustc_const_unstable(feature = "const_iter", issue = "92476")]
+impl<A: [const] Step + [const] Destruct> const DoubleEndedIterator for ops::Range<A> {


Is this change needed for the MVP? should at least be a separate commit unless the rest of this commit requires it

View changes since the review

It is, in fact, required since last() is implemented using next_back(), so it needs to be constified too.

fee1-dead · 2026-05-08T16:55:56Z

I feel iffy on introducing additional const impls that use specialization without us fixing specialization first

oli-obk · 2026-05-13T14:28:34Z

@rustbot blocked #156454

Filiprogrammer · 2026-06-20T16:05:02Z

This no longer builds since bc2112e.

error: expected a trait, found type
   --> /rustc/1df360b6a885af638e597c167cbb537b3b2137cd/library/core/src/iter/range.rs:819:42
    |
819 | impl<A: [const] Step + [const] Destruct> const RangeIteratorImpl for ops::Range<A> {
    |                                          ^^^^^^^^^^^^^^^^^^^^^^^

error: expected a trait, found type
   --> /rustc/1df360b6a885af638e597c167cbb537b3b2137cd/library/core/src/iter/traits/iterator.rs:321:44
    |
321 |         impl<I: [const] Iterator + ?Sized> const SpecAdvanceBy for I
    |                                            ^^^^^^^^^^^^^^^^^^^

Probably related to a88521a.

Randl · 2026-06-20T17:10:10Z

Looks like genuine ICE for me, I may take a look at it tomorrow

Randl · 2026-08-17T22:53:55Z

I'll rebase once I get to PC, but it is still blocked on const min specialization resolution (see discussion at https://rust-lang.zulipchat.com/#narrow/channel/326866-t-types.2Fnominated/topic/.23148200.3A.20min_specialization.20with.20const.20traits.20can.20call.20non-.E2.80.A6/with/614859233 )

I do think (and had a feeling that others agree) that this feature, i.e. allowing for in const, as well as constification of the Iterator trait, are well within constification policy and we probably want to play with both before stabilizing const traits

clarfonthey · 2026-08-18T00:46:26Z

To be clear, I mostly said what I did because I hadn't fully read the PR and didn't want to make any guarantees. Taking a second look, nothing looks particularly egregious, although will probably want a second opinion from wg-const-traits on what some of these tests were intended to do since it looks like they were relying on iterator for ranges not being implemented.

Randl · 2026-08-22T07:32:19Z

@rustbot ready

Rebased on master.
I checked the tests; I think the remaining ones should pass.
Still blocked on #156454 / whatever resolution of #148200 is

Randl · 2026-09-02T04:51:14Z

If we're okay with specializing on const (and not maybe-const), I think we can merge this.
Note that we can't just make TrustedLen non-const because that would lead to conflicting implementations of RangeIteratorImpl.
cc @clarfonthey @fee1-dead

JohnTitor · 2026-09-02T11:34:37Z

r? libs

clarfonthey · 2026-09-13T19:00:11Z

Returning from RustConf and finally looking at this: I think that the constification of the various methods on Iterator and DoubleEndedIterator should be done via a separate PR, although nothing in particular bothers me about that implementation.

(Feel free to r? clarfonthey on that PR.)

In terms of the rest, I'm a bit worried about the UI tests being affected, since I don't think the newer versions of the tests are testing what we want to test. The codegen test, however, looks fine, since it's an expected equivalence created by switching to a for loop.

I'm going to mark this as blocked until a separate PR for the iterator changes, but like I said, would be happy to review that separately if you submit it, and you can specify me for review. It's mostly just so that we can separate these out for loads of reasons, including potential perf checks + to get a bit more scrutiny on what exactly is being affected by that versus the range changes.

@rustbot blocked

Randl · 2026-09-13T19:16:38Z

    let mut x = 0;

    for i in 0..4 {
-        //~^ ERROR: cannot use `for`


This seems to clearly test loops in const context

View changes since the review

Randl · 2026-09-13T19:17:15Z


 const _: () = {
    for _ in 0..5 {}
-    //~^ ERROR cannot use `for`


this also appears to test specifically for for in const, though it duplicates

View changes since the review

Randl · 2026-09-13T19:34:04Z

I've opened #162734 for Iterator and DoubleEndedIterator methods.
As for tests, two are now passing. Both seem to test specifically the ability to use for in const. Some additional tests now have extra error messages related to feature gates for parts of the code that are potentially working on nightly now but still testing the intended behavior, as far as I understand.

Constify more Iterator and DoubleEndedIterator functions required for Range Extracted from rust-lang#156216 Consitfies functions required for `const impl Iterator for Range`. r? clarfonthey

Rollup merge of #162734 - Randl:const-for-prereq, r=clarfonthey Constify more Iterator and DoubleEndedIterator functions required for Range Extracted from #156216 Consitfies functions required for `const impl Iterator for Range`. r? clarfonthey

rustbot · 2026-09-14T07:43:23Z

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.

rustbot assigned oli-obk May 6, 2026

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. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels May 6, 2026

This comment has been minimized.

Sign in to view

Randl force-pushed the const-for-impl branch from d4ee14e to 6115579 Compare May 6, 2026 02:08

This comment has been minimized.

Sign in to view

Randl force-pushed the const-for-impl branch from 6115579 to 58f0c19 Compare May 6, 2026 04:23

theemathas reviewed May 6, 2026

View reviewed changes

Comment thread library/core/src/tuple.rs Outdated

Randl force-pushed the const-for-impl branch from 58f0c19 to ccbb26c Compare May 6, 2026 08:18

oli-obk reviewed May 6, 2026

View reviewed changes

rustbot added S-blocked Status: Blocked on something else such as an RFC or other implementation work. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels May 13, 2026

oli-obk reviewed Jun 20, 2026

View reviewed changes

Comment thread library/core/src/iter/range.rs Outdated

Randl force-pushed the const-for-impl branch from ccbb26c to b5015b7 Compare June 20, 2026 16:49

This comment has been minimized.

Sign in to view

Randl mentioned this pull request Jun 20, 2026

[ICE]: next solver const FnOnce bound on enum variant constructor hits fn_is_const debug assertion #158188

Closed

Randl force-pushed the const-for-impl branch from b5015b7 to 63fc647 Compare June 21, 2026 18:49

This comment has been minimized.

Sign in to view

Randl mentioned this pull request Aug 9, 2026

constify comparison traits on sliced types #147790

Open

This comment has been minimized.

Sign in to view

Randl force-pushed the const-for-impl branch from 63fc647 to b39ec5d Compare August 22, 2026 07:32

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 Aug 22, 2026

Randl force-pushed the const-for-impl branch from b39ec5d to 6d992a8 Compare September 2, 2026 04:41

This comment has been minimized.

Sign in to view

rustbot assigned clarfonthey and unassigned JohnTitor Sep 2, 2026

rustbot added S-blocked Status: Blocked on something else such as an RFC or other implementation work. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 13, 2026

Randl commented Sep 13, 2026

View reviewed changes

Randl mentioned this pull request Sep 13, 2026

Constify more Iterator and DoubleEndedIterator functions required for Range #162734

Merged

This comment has been minimized.

Sign in to view

Randl added 5 commits September 14, 2026 10:41

Constify more Iterator functions

2f9fea1

implement const Iterator for Range

7b633cc

Revert some const hacks

0ce5394

Fix tests

8179204

Don't specialize on maybe-const

dac6393

Randl force-pushed the const-for-impl branch from 6d992a8 to dac6393 Compare September 14, 2026 07:43

Uh oh!

Uh oh!

Conversation

Randl commented May 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Randl commented May 6, 2026

Uh oh!

This comment has been minimized.

theemathas commented May 6, 2026

Uh oh!

This comment has been minimized.

Randl commented May 6, 2026

Uh oh!

theemathas commented May 6, 2026

Uh oh!

Uh oh!

oli-obk May 6, 2026 • 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!

Randl May 8, 2026

Choose a reason for hiding this comment

Uh oh!

fee1-dead commented May 8, 2026

Uh oh!

oli-obk commented May 13, 2026

Uh oh!

Filiprogrammer commented Jun 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

Randl commented Jun 20, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

Randl commented Aug 17, 2026

Uh oh!

clarfonthey commented Aug 18, 2026

Uh oh!

This comment has been minimized.

Randl commented Aug 22, 2026

Uh oh!

This comment has been minimized.

Randl commented Sep 2, 2026

Uh oh!

JohnTitor commented Sep 2, 2026

Uh oh!

clarfonthey commented Sep 13, 2026

Uh oh!

Randl Sep 13, 2026 • 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!

Randl Sep 13, 2026 • 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!

Randl commented Sep 13, 2026

Uh oh!

This comment has been minimized.

rustbot commented Sep 14, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Randl commented May 6, 2026 •

edited by rustbot

Loading

oli-obk May 6, 2026 •

edited by rustbot

Loading

Filiprogrammer commented Jun 20, 2026 •

edited

Loading

Randl Sep 13, 2026 •

edited by rustbot

Loading

Randl Sep 13, 2026 •

edited by rustbot

Loading