Raw handle macros for field projection experiments by dingxiangfei2009 · Pull Request #164627 · rust-lang/rust

dingxiangfei2009 · 2026-10-20T09:39:14Z

🐉 Highly experimental field projection primitive `raw_handle!`

We would like to submit this patch to start the experiment with the LocalHandle and PlaceHandle. We intend to continue polishing it in a patch series following this initial patch.

We know that the following items are not working.

Closure capture is still hanging in the air, this will be another major work.
Library trait wording. Field projection design in general is still shifting landscape, so many pieces as presented here would change. We will continue submitting patches to match the design.

Signed-off-by: Xiangfei Ding <dingxiangfei2009@protonmail.ch>

dingxiangfei2009 · 2026-11-09T15:15:33Z

@rustbot ready

rustbot · 2026-11-09T15:23:11Z

Some changes occurred in compiler/rustc_attr_ir

cc @jdonszelmann, @JonathanBrouwer

clippy is developed in its own repository. If possible, consider making this change to rust-lang/rust-clippy instead.

cc @rust-lang/clippy

HIR ty lowering was modified

cc @fmease

rustfmt is developed in its own repository. If possible, consider making this change to rust-lang/rustfmt instead.

cc @rust-lang/rustfmt

The parser was modified, potentially altering the grammar of (stable) Rust
which would be a breaking change.

cc @fmease

BennoLossin · 2026-11-09T15:33:25Z

hey @oli-obk it would be great if we could get this merged before the cutoff for 1.102. we have some further additions that we'll want to make, but those can be done in separate PRs. I'd also really like to get to a more incremental dev cycle here.

one note on this PR, Ding and I were in a meeting just now and we noticed that we don't yet have a Deref variant in FieldProjectKind. Ding will add that later today.

Co-authored-by: Xiangfei Ding <dingxiangfei2009protonmail.ch> Signed-off-by: Rasmus Piorr <piorr@google.com> Signed-off-by: Xiangfei Ding <dingxiangfei2009@protonmail.ch>

rust-bors · 2026-11-10T20:31:24Z

☔ The latest upstream changes (presumably #165760) made this pull request unmergeable. Please resolve the merge conflicts by rebasing.

oli-obk

First review round

View changes since this review

oli-obk · 2026-11-11T09:32:07Z

+                let place_with_id = self.cat_res(opr.hir_id, opr.span, root_ty, res)?;
+                // NOTE(dingxiangfei2009,shaddydc): Technically we should capture the root
+                // local **via LocalHandle**, but upvar does not understand this yet.
+                // We shall come back and add a new upvar capture just for this.


what is the effect of the current capture system? overcapturing? document the current behaviour and what changing it would mean

oli-obk · 2026-11-11T09:33:37Z

        }
    }

+    fn lower_raw_handle_step(


move these functions and everything else field projection related into a seprate module

oli-obk · 2026-11-11T09:36:44Z

 pub macro raw_handle($place:expr) {
    k#raw_handle($place)
 }


btw, why does this go through the custom syntax scheme instead of just having a builtin macro expand directly to the right ast without having the parser handle it?

As per MCP 580 | Add builtin# for compiler-intrinsic syntax, built-in macros should not expand to syntactic constructs that cannot be otherwise named in the surface language since it would blur the line between macros and "language extensions" as the MCP calls it.

So for example a built-in macro like line is fine because it it expands to a string literal which can be named in the surface language.

Expanding to internal syntax isn't necessarily done to write less code in the compiler, it's mainly just a matter of principle.

oli-obk · 2026-11-11T09:40:12Z

    OffsetOf(&'hir Ty<'hir>, &'hir [Ident]),

+    /// Raw place handle (`raw_handle!`)
+    RawHandle(&'hir FieldProject<'hir>),


do you need this as a hir::Node anywhere? It's always part of an expression, so it could just not have its own HirId and just live in the ExprKind

oli-obk · 2026-11-11T09:41:23Z

                self.pclose();
            }
+            ast::ExprKind::RawHandle(expr) => {
+                self.word("k#raw_handle");


may be better to print as the macro, and probably should add a rustfmt test now, too, otherwise you may get funny things like rustfmt changing your raw_handle!(...) call into k#raw_handle(...)

oli-obk · 2026-11-11T09:43:29Z

    fn visit_expr(&mut self, ex: &'v Expr<'v>) -> Self::Result {
        walk_expr(self, ex)
    }
+    fn visit_field_project(&mut self, fp: &'v FieldProject<'v>) -> Self::Result {


do you need this for anything? Could only add it once needed

oli-obk · 2026-11-11T09:44:16Z

+#[derive(Debug, Clone, Copy, StableHash)]
+pub struct FieldProject<'hir> {
+    #[stable_hash(ignore)]
+    pub hir_id: HirId,


Don't really need this id afaict, the surrounding expression always already has an id. Or is this its own node because you wanna store sth per nested FieldProject?

I wonder if we should just have prefix: &[FieldProjectKind] and path: QPath fields instead of having a nested enum structure that always terminates with Path.

rustbot · 2026-11-11T09:48:07Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

rustbot assigned oli-obk Oct 20, 2026

dingxiangfei2009 force-pushed the raw-handle-macros branch from 4cdb718 to d5bad9f Compare October 20, 2026 09:39

This comment has been minimized.

Sign in to view

ojeda mentioned this pull request Oct 20, 2026

Rust unstable features needed for the kernel Rust-for-Linux/linux#2

Open

57 tasks

BennoLossin reviewed Oct 20, 2026

View reviewed changes

Comment thread compiler/rustc_ast/src/ast.rs Resolved

dingxiangfei2009 force-pushed the raw-handle-macros branch from d5bad9f to a0c239f Compare October 21, 2026 11:26

This comment has been minimized.

Sign in to view

add core::ops::place skeleton

d6ac639

Signed-off-by: Xiangfei Ding <dingxiangfei2009@protonmail.ch>

dingxiangfei2009 force-pushed the raw-handle-macros branch from a0c239f to 33785a7 Compare November 9, 2026 14:44

oli-obk added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. A-repr Area: the `#[repr(stuff)]` attribute and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Nov 9, 2026

dingxiangfei2009 marked this pull request as ready for review November 9, 2026 15:23

dingxiangfei2009 force-pushed the raw-handle-macros branch from 33785a7 to 7815525 Compare November 9, 2026 21:56

This comment has been minimized.

Sign in to view

dingxiangfei2009 force-pushed the raw-handle-macros branch 3 times, most recently from bd6ca0b to aeca9e4 Compare November 10, 2026 14:38

implement k# raw_handle and place handles

fddb240

Co-authored-by: Xiangfei Ding <dingxiangfei2009protonmail.ch> Signed-off-by: Rasmus Piorr <piorr@google.com> Signed-off-by: Xiangfei Ding <dingxiangfei2009@protonmail.ch>

dingxiangfei2009 force-pushed the raw-handle-macros branch from aeca9e4 to fddb240 Compare November 10, 2026 14:52

oli-obk requested changes Nov 11, 2026

View reviewed changes

rustbot 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 Nov 11, 2026

Conversation

dingxiangfei2009 commented Oct 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🐉 Highly experimental field projection primitive raw_handle!

Uh oh!

This comment has been minimized.

Uh oh!

This comment has been minimized.

This comment has been minimized.

dingxiangfei2009 commented Nov 9, 2026

Uh oh!

rustbot commented Nov 9, 2026

Uh oh!

BennoLossin commented Nov 9, 2026

Uh oh!

This comment has been minimized.

rust-bors Bot commented Nov 10, 2026

Uh oh!

oli-obk 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!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

fmease Nov 13, 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!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

oli-obk Nov 11, 2026

Choose a reason for hiding this comment

Uh oh!

rustbot commented Nov 11, 2026

Uh oh!

This branch has not been deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

dingxiangfei2009 commented Oct 20, 2026 •

edited

Loading

🐉 Highly experimental field projection primitive `raw_handle!`

oli-obk left a comment •

edited by rustbot

Loading

fmease Nov 13, 2026 •

edited

Loading