Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List` by spirali · Pull Request #152593 · rust-lang/rust

spirali · 2026-02-13T23:26:24Z

This is related to trait system refactoring. It fixes the FIXME in ValTreeKind

   // FIXME(mgca): Use a `List` here instead of a boxed slice
    Branch(Box<[I::Const]>),

It introduces Interner::Consts, changes Branch(Box<[I::Const]>) to Branch(I::Consts), and updates all relevant places.

r? lcnr

rustbot · 2026-02-13T23:26:29Z

Some changes occurred in match lowering

cc @Nadrieril

lcnr · 2026-02-14T17:46:24Z

compiler/rustc_mir_build/src/builder/matches/mod.rs

+                let ValTreeKind::Branch(branch) = *valtree else {
+                    bug!("malformed valtree for an enum")
+                };
+                let Some(actual_variant_idx) = branch.get(0) else {


we stop asserting that the len is 1 here

lcnr · 2026-02-14T17:48:41Z

compiler/rustc_type_ir/src/const_kind.rs

@@ -179,7 +179,7 @@ impl<I: Interner> ValTreeKind<I> {
    #[inline]
    pub fn to_branch(&self) -> &[I::Const] {


Suggested change

pub fn to_branch(&self) -> &[I::Const] {

pub fn to_branch(&self) -> I::Consts {

lcnr · 2026-02-14T17:48:48Z

compiler/rustc_type_ir/src/const_kind.rs

@@ -195,7 +195,7 @@ impl<I: Interner> ValTreeKind<I> {
    /// Attempts to convert to a `ValTreeKind::Branch` value.
    pub fn try_to_branch(&self) -> Option<&[I::Const]> {


lcnr

please remove the fn kind method from trait ValTree and instead implement IntoKind for it now.

VaTreeKind can now be Copy

View changes since this review

lcnr · 2026-02-14T17:52:05Z

compiler/rustc_middle/src/ty/print/pretty.rs

            }
            // Otherwise, print the array separated by commas (or if it's a tuple)
            (ty::ValTreeKind::Branch(fields), ty::Array(..) | ty::Tuple(..)) => {
-                let fields_iter = fields.iter().copied();


Consts should implement SliceLike so we can just do .iter() here now

Consts implements SliceLike, but .iter() on fields returns &Const and the following code works only for iterator over Const:

error[E0277]: the trait bound `&ty::consts::Const<'_>: print::Print<'tcx, Self>` is not satisfied --> compiler/rustc_middle/src/ty/print/pretty.rs:1964:34 | 1964 | ... self.comma_sep(fields)?; | ^^^^^^^^^ the nightly-only, unstable trait `print::Print<'tcx, Self>` is not implemented for `&ty::consts::Const<'_>` | help: the trait `Print<'tcx, _>` is not implemented for `&ty::consts::Const<'_>` but trait `Print<'_, _>` is implemented for `ty::consts::Const<'_>

I do think the as_slice should not be needed 🤔

Though given that we're in rustc_middle i guess this is the inherent impl for List or via deref?

still relevant, can you try removing the as_slice?

^ (get rid of as_slice)

Ah, sorry, I have overlooked your original reply. It is fixed now.

lcnr · 2026-02-14T17:54:54Z

compiler/rustc_type_ir/src/relate.rs

+                        .as_slice()
+                        .iter()
+                        .zip(branches_b.as_slice().iter())
                        .all(|(a, b)| relation.relate(*a, *b).is_ok())


Suggested change

.as_slice()

.iter()

.zip(branches_b.as_slice().iter())

.all(|(a, b)| relation.relate(*a, *b).is_ok())

.iter()

.zip(branches_b.iter())

.all(|(a, b)| relation.relate(a, b).is_ok())

spirali · 2026-02-15T08:15:32Z

please remove the fn kind method from trait ValTree and instead implement IntoKind for it now.

VaTreeKind can now be Copy

View changes since this review

When ValTreeKind is Copy, I get the following error:

error[E0283]: type annotations needed
   --> compiler/rustc_middle/src/ty/context.rs:564:79
    |
564 |                 interners.valtree.intern(v, |v| InternedInSet(interners.arena.alloc(v))).0,
    |                                                                               ^^^^^ cannot infer type of the type parameter `C` declared on the method `alloc`
    |
note: multiple `impl`s satisfying `rustc_type_ir::ValTreeKind<context::TyCtxt<'tcx>>: arena::ArenaAllocatable<'_, _>` found
   --> compiler/rustc_arena/src/lib.rs:646:5
    |
628 |    pub macro declare_arena([$($a:tt $name:ident: $ty:ty,)*]) {
    |    ----------------------- in this expansion of `rustc_arena::declare_arena!` (#2)
...
646 |        impl<'tcx, T: Copy> ArenaAllocatable<'tcx, rustc_arena::IsCopy> for T {
    |        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...
662 |            impl<'tcx> ArenaAllocatable<'tcx, rustc_arena::IsNotCopy> for $ty {
    |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

If I try to fix it through adding explicit type:

InternedInSet(self.interners.arena.alloc::<_, rustc_arena::IsCopy>(v))

I get:

error[E0277]: the trait bound `Allocation: arena::ArenaAllocatable<'_, IsCopy>` is not satisfied

I do not understand what is the problem here.

lcnr · 2026-02-15T12:28:05Z

Can you try removing the ValTreeKind from

rust/compiler/rustc_middle/src/arena.rs

Line 95 in ce0bf0b

[] valtree: rustc_middle::ty::ValTreeKind<rustc_middle::ty::TyCtxt<'tcx>>,

valtrees can now be stored in the dropless arena.

I don't 100% know how to do here and you could look at how e.g. Ty<'tcx> arena allocation is handled

rustbot · 2026-02-17T20:44:07Z

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

spirali · 2026-02-17T20:52:44Z

Can you try removing the ValTreeKind from

rust/compiler/rustc_middle/src/arena.rs

Line 95 in ce0bf0b

[] valtree: rustc_middle::ty::ValTreeKind<rustc_middle::ty::TyCtxt<'tcx>>,

valtrees can now be stored in the dropless arena.

I don't 100% know how to do here and you could look at how e.g. Ty<'tcx> arena allocation is handled

Thank you, it is fixed now.

ValTreeKind is now copy, but in the following test returns a different hash:

static fn function_names::const_generic_fn_non_int<{CONST#ffa3db4ca1d52dce}>();

With this change, it is safe to update the test to a new hash value or I would just cover a deeper problem?

lcnr · 2026-02-22T20:31:50Z

With this change, it is safe to update the test to a new hash value or I would just cover a deeper problem?

I think this is fine, but not 100% sure 🤔 feel free to bless it and I then look at the change before merge

@bors try @rust-timer queue

Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List`

rust-bors · 2026-02-22T22:43:33Z

☀️ Try build successful (CI)
Build commit: 7e95d4f (7e95d4f9eee300952b21051884d058f1f5bddea3, parent: 1500f0f47f5fe8ddcd6528f6c6c031b210b4eac5)

rust-timer · 2026-02-22T23:47:40Z

Finished benchmarking commit (7e95d4f): comparison URL.

Overall result: no relevant changes - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results (primary -0.4%, secondary 0.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	1.9%	[1.5%, 2.2%]	2
Regressions ❌ (secondary)	3.0%	[3.0%, 3.0%]	1
Improvements ✅ (primary)	-1.9%	[-3.0%, -0.9%]	3
Improvements ✅ (secondary)	-1.1%	[-1.1%, -1.1%]	2
All ❌✅ (primary)	-0.4%	[-3.0%, 2.2%]	5

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 483.037s -> 480.952s (-0.43%)
Artifact size: 397.95 MiB -> 398.00 MiB (0.01%)

lcnr · 2026-03-02T07:33:13Z

tests/debuginfo/function-names.rs

 //@ gdb-command:info functions -q function_names::const_generic_fn.*
 //@ gdb-check:[...]static fn function_names::const_generic_fn_bool<false>();
-//@ gdb-check:[...]static fn function_names::const_generic_fn_non_int<{CONST#ffa3db4ca1d52dce}>();
+//@ gdb-check:[...]static fn function_names::const_generic_fn_non_int<{CONST#6dd80cc0c950c171}>();


that's fine, different HashStable impl between List and boxed slices 🤷

Thanks. Does it need some additional work from my side? All comments above should be resolved

lcnr

r=me after that nit

View changes since this review

lcnr · 2026-03-05T09:35:38Z

compiler/rustc_middle/src/ty/print/pretty.rs

            }
            // Otherwise, print the array separated by commas (or if it's a tuple)
            (ty::ValTreeKind::Branch(fields), ty::Array(..) | ty::Tuple(..)) => {
-                let fields_iter = fields.iter().copied();


^ (get rid of as_slice)

Kobzol · 2026-03-05T14:01:42Z

@bors delegate

rust-bors · 2026-03-05T14:01:45Z

✌️ @spirali, you can now approve this pull request!

If @lcnr told you to "r=me" after making some further change, then please make that change and post @bors r=lcnr.

spirali · 2026-03-05T21:50:42Z

@bors r=lcnr

rust-bors · 2026-03-05T21:50:45Z

📌 Commit f8998e9 has been approved by lcnr

It is now in the queue for this repository.

Kobzol · 2026-03-05T21:53:29Z

The perf. result was neutral on the last perf. run, so:

@bors rollup=maybe

Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List` This is related to trait system refactoring. It fixes the FIXME in `ValTreeKind` ``` // FIXME(mgca): Use a `List` here instead of a boxed slice Branch(Box<[I::Const]>), ``` It introduces `Internator::Consts`, changes `Branch(Box<[I::Const]>)` to `Branch(I::Consts)`, and updates all relevant places. r? lcnr

Rollup of 7 pull requests Successful merges: - #153466 (`rust-analyzer` subtree update) - #151280 (Fix incorrect trailing comma suggested in no_accessible_fields) - #152593 (Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List`) - #153189 (refactor: move `check_align` to `parse_alignment`) - #153230 (Roll rustfmt reviewers for in-tree rustfmt) - #153321 (Add high-priority ICEs to tests/crashes) - #153445 (Consider try blocks as block-like for overflowed expr)

Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List` This is related to trait system refactoring. It fixes the FIXME in `ValTreeKind` ``` // FIXME(mgca): Use a `List` here instead of a boxed slice Branch(Box<[I::Const]>), ``` It introduces `Interner::Consts`, changes `Branch(Box<[I::Const]>)` to `Branch(I::Consts)`, and updates all relevant places. r? lcnr

…uwer Rollup of 9 pull requests Successful merges: - #153466 (`rust-analyzer` subtree update) - #151280 (Fix incorrect trailing comma suggested in no_accessible_fields) - #152593 (Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List`) - #153174 (std: add wasm64 to sync::Once and thread_parking atomics cfg guards) - #153189 (refactor: move `check_align` to `parse_alignment`) - #153230 (Roll rustfmt reviewers for in-tree rustfmt) - #153445 (Consider try blocks as block-like for overflowed expr) - #153476 (bootstrap.py: fix typo "parallle") - #153483 (Preserve parentheses around `Fn` trait bounds in pretty printer)

Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List` This is related to trait system refactoring. It fixes the FIXME in `ValTreeKind` ``` // FIXME(mgca): Use a `List` here instead of a boxed slice Branch(Box<[I::Const]>), ``` It introduces `Interner::Consts`, changes `Branch(Box<[I::Const]>)` to `Branch(I::Consts)`, and updates all relevant places. r? lcnr

…uwer Rollup of 14 pull requests Successful merges: - #153466 (`rust-analyzer` subtree update) - #151280 (Fix incorrect trailing comma suggested in no_accessible_fields) - #152593 (Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List`) - #153174 (std: add wasm64 to sync::Once and thread_parking atomics cfg guards) - #153485 (libcore float tests: replace macro shadowing by const-compatible macro) - #153495 (Fix ICE in `offset_of!` error recovery) - #152040 (Do not emit ConstEvaluatable goals if type-const) - #152741 (Suppress invalid suggestions in destructuring assignment) - #153189 (refactor: move `check_align` to `parse_alignment`) - #153230 (Roll rustfmt reviewers for in-tree rustfmt) - #153445 (Consider try blocks as block-like for overflowed expr) - #153452 (Cleanup unused diagnostic emission methods) - #153476 (bootstrap.py: fix typo "parallle") - #153483 (Preserve parentheses around `Fn` trait bounds in pretty printer)

Rollup merge of #152593 - spirali:valtreekind-list, r=lcnr Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List` This is related to trait system refactoring. It fixes the FIXME in `ValTreeKind` ``` // FIXME(mgca): Use a `List` here instead of a boxed slice Branch(Box<[I::Const]>), ``` It introduces `Interner::Consts`, changes `Branch(Box<[I::Const]>)` to `Branch(I::Consts)`, and updates all relevant places. r? lcnr

…uwer Rollup of 14 pull requests Successful merges: - rust-lang/rust#153466 (`rust-analyzer` subtree update) - rust-lang/rust#151280 (Fix incorrect trailing comma suggested in no_accessible_fields) - rust-lang/rust#152593 (Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List`) - rust-lang/rust#153174 (std: add wasm64 to sync::Once and thread_parking atomics cfg guards) - rust-lang/rust#153485 (libcore float tests: replace macro shadowing by const-compatible macro) - rust-lang/rust#153495 (Fix ICE in `offset_of!` error recovery) - rust-lang/rust#152040 (Do not emit ConstEvaluatable goals if type-const) - rust-lang/rust#152741 (Suppress invalid suggestions in destructuring assignment) - rust-lang/rust#153189 (refactor: move `check_align` to `parse_alignment`) - rust-lang/rust#153230 (Roll rustfmt reviewers for in-tree rustfmt) - rust-lang/rust#153445 (Consider try blocks as block-like for overflowed expr) - rust-lang/rust#153452 (Cleanup unused diagnostic emission methods) - rust-lang/rust#153476 (bootstrap.py: fix typo "parallle") - rust-lang/rust#153483 (Preserve parentheses around `Fn` trait bounds in pretty printer)

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 Feb 13, 2026

rustbot assigned lcnr Feb 13, 2026

ValTreeKind::Branch use List

0f786b9

spirali force-pushed the valtreekind-list branch from 721e973 to 0f786b9 Compare February 14, 2026 09:59

This comment has been minimized.

Sign in to view

lcnr reviewed Feb 14, 2026

View reviewed changes

Changes after code review

46ddd48

This comment has been minimized.

Sign in to view

rust-bors bot pushed a commit that referenced this pull request Feb 22, 2026

Auto merge of #152593 - spirali:valtreekind-list, r=<try>

7e95d4f

Box in `ValTreeKind::Branch(Box<[I::Const]>)` changed to `List`

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Feb 22, 2026

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Feb 22, 2026

Test updated

14b3a40

lcnr reviewed Mar 2, 2026

View reviewed changes

lcnr reviewed Mar 5, 2026

View reviewed changes

as_slice removed

f8998e9

rust-bors bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Mar 5, 2026

jhpratt mentioned this pull request Mar 6, 2026

Rollup of 7 pull requests #153467

Closed

JonathanBrouwer mentioned this pull request Mar 6, 2026

Rollup of 9 pull requests #153490

Closed

This was referenced Mar 6, 2026

Rollup of 13 pull requests #153498

Closed

Rollup of 14 pull requests #153507

Merged

rust-bors bot merged commit 77658bc into rust-lang:main Mar 6, 2026
11 checks passed

rustbot added this to the 1.96.0 milestone Mar 6, 2026

		@@ -179,7 +179,7 @@ impl<I: Interner> ValTreeKind<I> {
		#[inline]
		pub fn to_branch(&self) -> &[I::Const] {

	pub fn to_branch(&self) -> &[I::Const] {
	pub fn to_branch(&self) -> I::Consts {

		@@ -195,7 +195,7 @@ impl<I: Interner> ValTreeKind<I> {
		/// Attempts to convert to a `ValTreeKind::Branch` value.
		pub fn try_to_branch(&self) -> Option<&[I::Const]> {

Uh oh!

Conversation

spirali commented Feb 13, 2026 • edited by lcnr Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Feb 13, 2026

Uh oh!

This comment has been minimized.

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!

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!

lcnr Mar 5, 2026 • edited 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!

spirali commented Feb 15, 2026

Uh oh!

lcnr commented Feb 15, 2026

Uh oh!

rustbot commented Feb 17, 2026

Uh oh!

spirali commented Feb 17, 2026

Uh oh!

This comment has been minimized.

lcnr commented Feb 22, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors bot commented Feb 22, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Feb 22, 2026

Overall result: no relevant changes - no action needed

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

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!

lcnr Mar 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Kobzol commented Mar 5, 2026

Uh oh!

rust-bors bot commented Mar 5, 2026 • edited by Kobzol Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

spirali commented Mar 5, 2026

Uh oh!

rust-bors bot commented Mar 5, 2026

Uh oh!

Kobzol commented Mar 5, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

spirali commented Feb 13, 2026 •

edited by lcnr

Loading

lcnr left a comment •

edited by rustbot

Loading

lcnr Mar 5, 2026 •

edited

Loading

lcnr left a comment •

edited by rustbot

Loading

lcnr Mar 5, 2026 •

edited

Loading

rust-bors bot commented Mar 5, 2026 •

edited by Kobzol

Loading