Lower multi-segment const paths as `ConstArgKind::Path` #135186

camelid · 2025-01-07T05:21:38Z

This PR:

Lower all const paths in arg position to hir::ConstArgKind::Path
Then lower assoc const paths to ty::ConstKind::Unevaluated
(incomplete) Introduce #[type_const] attribute for trait assoc consts that are allowed as const args

Next steps:

Implement code to check that assoc const definitions satisfy #[type_const] if present (basically is it a const path or monomorphic anon const)

r? @BoxyUwU

rustbot · 2025-01-11T02:51:48Z

Some changes occurred in const_evaluatable.rs

cc @BoxyUwU

HIR ty lowering was modified

cc @fmease

BoxyUwU

thx for working on this :3

BoxyUwU · 2025-01-11T03:15:01Z

compiler/rustc_ast_lowering/src/lib.rs

@@ -1100,7 +1100,8 @@ impl<'a, 'hir> LoweringContext<'a, 'hir> {
                            .and_then(|partial_res| partial_res.full_res())
                        {
                            if !res.matches_ns(Namespace::TypeNS)
-                                && path.is_potential_trivial_const_arg()
+                            // FIXME: should this only allow single-segment paths?


Suggested change

// FIXME: should this only allow single-segment paths?

// FIXME(mgca): should this only allow single-segment paths?

unannotated FIXME will just get lost tbh, no way of tracking them down lol

Ah yeah, I meant it as a FIXME to deal with before merging this PR -- which is why I didn't label it. But since all of this is experimental anyway, we could deal with it later.

Oh in that case using // TODO will fail tidy and ensure the PR can't be merged until its resolved which ensures noone can forget about it before r+

BoxyUwU · 2025-01-14T18:01:22Z

compiler/rustc_ast/src/ast.rs

    /// Could this expr be either `N`, or `{ N }`, where `N` is a const parameter.
    ///
    /// If this is not the case, name resolution does not resolve `N` when using
    /// `min_const_generics` as more complex expressions are not supported.
    ///
    /// Does not ensure that the path resolves to a const param, the caller should check this.
    /// This also does not consider macros, so it's only correct after macro-expansion.
-    pub fn is_potential_trivial_const_arg(&self) -> bool {
+    pub fn is_potential_trivial_const_arg(&self, allow_multi_segment: bool) -> bool {


It's not that just we should support multi seg paths we should also support generic arguments on paths too, and also support qualified paths. e.g. this path which has segments with args and also a self ty: <T as Trait<u32>>::ASSOC

BoxyUwU · 2025-01-16T05:32:03Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

-    /// Literals and const generic parameters are eagerly converted to a constant, everything else
-    /// becomes `Unevaluated`.
+    #[instrument(level = "debug", skip(self), ret)]
+    pub fn lower_const_assoc_path(


We should definitely consolidate this and lower_assoc_path into one method that handles 99% of the logic since the only real behavioural differences are the AssocKind arguments we hard code and that we construct a Ty or Const at the very end depending on which method it is

BoxyUwU · 2025-01-16T05:32:27Z

compiler/rustc_hir_analysis/src/hir_ty_lowering/mod.rs

+            match &candidates[..] {
+                [] => {}
+                &[(impl_, assoc)] => {
+                    // FIXME(mgca): adapted from temporary inherent assoc ty code that may be incorrect


same here we should just make probe_inherent_assoc_ty be able to be reused for consts somehow

BoxyUwU · 2025-01-16T05:33:38Z

compiler/rustc_hir_analysis/src/collect.rs

@@ -549,6 +549,95 @@ impl<'tcx> HirTyLowerer<'tcx> for ItemCtxt<'tcx> {
        }
    }

+    fn lower_assoc_const(


this would probably also be good to unify with lower_assoc_ty somehow, at the very least it would be good to not duplicate almost 100 lines of diagnostics logic :3

rust-log-analyzer · 2025-01-20T06:13:24Z

The job x86_64-gnu-llvm-18 failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

#21 exporting to docker image format
#21 sending tarball 28.4s done
#21 DONE 34.3s
##[endgroup]
Setting extra environment values for docker:  --env ENABLE_GCC_CODEGEN=1 --env GCC_EXEC_PREFIX=/usr/lib/gcc/
[CI_JOB_NAME=x86_64-gnu-llvm-18]
debug: `DISABLE_CI_RUSTC_IF_INCOMPATIBLE` configured.
---
sccache: Starting the server...
##[group]Configure the build
configure: processing command line
configure: 
configure: build.configure-args := ['--build=x86_64-unknown-linux-gnu', '--llvm-root=/usr/lib/llvm-18', '--enable-llvm-link-shared', '--set', 'rust.randomize-layout=true', '--set', 'rust.thin-lto-import-instr-limit=10', '--enable-verbose-configure', '--enable-sccache', '--disable-manage-submodules', '--enable-locked-deps', '--enable-cargo-native-static', '--set', 'rust.codegen-units-std=1', '--set', 'dist.compression-profile=balanced', '--dist-compression-formats=xz', '--set', 'rust.lld=false', '--disable-dist-src', '--release-channel=nightly', '--enable-debug-assertions', '--enable-overflow-checks', '--enable-llvm-assertions', '--set', 'rust.verify-llvm-ir', '--set', 'rust.codegen-backends=llvm,cranelift,gcc', '--set', 'llvm.static-libstdcpp', '--enable-new-symbol-mangling']
configure: target.x86_64-unknown-linux-gnu.llvm-config := /usr/lib/llvm-18/bin/llvm-config
configure: llvm.link-shared     := True
configure: rust.randomize-layout := True
configure: rust.thin-lto-import-instr-limit := 10
---
##[group]Testing stage2 compiletest suite=ui mode=ui (x86_64-unknown-linux-gnu)

running 18301 tests
10%  --  1831/18301,  1786 passed, 0 failed, 45 ignored
   [ui] tests/ui/const-generics/mgca/assoc-const.rs ... FAILED
30%  --  5491/18301,  5429 passed, 1 failed, 61 ignored
40%  --  7321/18301,  7250 passed, 1 failed, 70 ignored
50%  --  9151/18301,  9068 passed, 1 failed, 82 ignored
60%  -- 10981/18301, 10870 passed, 1 failed, 110 ignored
---


failures:

---- [ui] tests/ui/const-generics/mgca/assoc-const.rs stdout ----
error: test compilation failed although it shouldn't!
status: exit status: 1
status: exit status: 1
command: env -u RUSTC_LOG_COLOR RUSTC_ICE="0" RUST_BACKTRACE="short" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2/bin/rustc" "/checkout/tests/ui/const-generics/mgca/assoc-const.rs" "-Zthreads=1" "-Zsimulate-remapped-rust-src-base=/rustc/FAKE_PREFIX" "-Ztranslate-remapped-path-to-local-path=no" "-Z" "ignore-directory-in-diagnostics-source-blocks=/cargo" "-Z" "ignore-directory-in-diagnostics-source-blocks=/checkout/vendor" "--sysroot" "/checkout/obj/build/x86_64-unknown-linux-gnu/stage2" "--target=x86_64-unknown-linux-gnu" "--check-cfg" "cfg(test,FALSE)" "--error-format" "json" "--json" "future-incompat" "-Ccodegen-units=1" "-Zui-testing" "-Zdeduplicate-diagnostics=no" "-Zwrite-long-types-to-disk=no" "-Cstrip=debuginfo" "--emit" "metadata" "-C" "prefer-dynamic" "--out-dir" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/const-generics/mgca/assoc-const" "-A" "unused" "-A" "internal_features" "-Crpath" "-Cdebuginfo=0" "-Lnative=/checkout/obj/build/x86_64-unknown-linux-gnu/native/rust-test-helpers"
--- stderr -------------------------------
error[E0277]: the trait bound `T: Tr<bool>` is not satisfied
error[E0277]: the trait bound `T: Tr<bool>` is not satisfied
##[error]  --> /checkout/tests/ui/const-generics/mgca/assoc-const.rs:11:42
   |
LL | fn mk_array<T: Tr<bool>>(_x: T) -> [(); <T as Tr<bool>>::SIZE] {
   |                                          ^ the trait `Tr<bool>` is not implemented for `T`
error: aborting due to 1 previous error

For more information about this error, try `rustc --explain E0277`.
------------------------------------------
------------------------------------------



failures:
    [ui] tests/ui/const-generics/mgca/assoc-const.rs
test result: FAILED. 18119 passed; 1 failed; 181 ignored; 0 measured; 0 filtered out; finished in 177.77s

Some tests failed in compiletest suite=ui mode=ui host=x86_64-unknown-linux-gnu target=x86_64-unknown-linux-gnu
Build completed unsuccessfully in 0:17:56

rustbot assigned BoxyUwU Jan 7, 2025

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 Jan 7, 2025

This comment has been minimized.

Sign in to view

camelid changed the title ~~Start lowering multi-segment const paths as ConstArgKind::Path~~ Lower multi-segment const paths as ConstArgKind::Path Jan 7, 2025

camelid force-pushed the const-path-multi branch from c501669 to 4efbed3 Compare January 7, 2025 18:52

This comment has been minimized.

Sign in to view

camelid force-pushed the const-path-multi branch from 4efbed3 to 0036c2c Compare January 7, 2025 19:13

This comment has been minimized.

Sign in to view

camelid marked this pull request as ready for review January 11, 2025 02:51

camelid changed the title ~~Lower multi-segment const paths as ConstArgKind::Path~~ Lower multi-segment const paths as ConstArgKind::Path Jan 11, 2025

BoxyUwU reviewed Jan 16, 2025

View reviewed changes

camelid added 8 commits January 19, 2025 20:47

Start lowering multi-segment const paths as ConstArgKind::Path

e4be4df

Partially fix parent_args handling for assoc const

45d7b04

Avoid "const uses param" error

7006146

Replace TODOs with FIXMEs to allow CI to pass for now

968cfaf

Properly lower assoc consts from traits

214fd98

Avoid "const depends on generics" error for bare paths under mgca

4f90c3b

Standardize mgca FIXMEs to "mgca"

eadca84

Start implementing #[type_const]

a296ff8

camelid force-pushed the const-path-multi branch from 263637d to a296ff8 Compare January 20, 2025 04:47

Allow const paths with generics/qself under mgca

8ffa35c

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Lower multi-segment const paths as `ConstArgKind::Path` #135186

Lower multi-segment const paths as `ConstArgKind::Path` #135186

camelid commented Jan 7, 2025 •

edited

Loading

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rustbot commented Jan 11, 2025

BoxyUwU left a comment

BoxyUwU Jan 11, 2025

camelid Jan 20, 2025

BoxyUwU Jan 20, 2025

BoxyUwU Jan 14, 2025

BoxyUwU Jan 16, 2025

BoxyUwU Jan 16, 2025

BoxyUwU Jan 16, 2025

rust-log-analyzer commented Jan 20, 2025

	// FIXME: should this only allow single-segment paths?
	// FIXME(mgca): should this only allow single-segment paths?

Lower multi-segment const paths as ConstArgKind::Path #135186

Are you sure you want to change the base?

Lower multi-segment const paths as ConstArgKind::Path #135186

Conversation

camelid commented Jan 7, 2025 • edited Loading

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rustbot commented Jan 11, 2025

BoxyUwU left a comment

Choose a reason for hiding this comment

BoxyUwU Jan 11, 2025

Choose a reason for hiding this comment

camelid Jan 20, 2025

Choose a reason for hiding this comment

BoxyUwU Jan 20, 2025

Choose a reason for hiding this comment

BoxyUwU Jan 14, 2025

Choose a reason for hiding this comment

BoxyUwU Jan 16, 2025

Choose a reason for hiding this comment

BoxyUwU Jan 16, 2025

Choose a reason for hiding this comment

BoxyUwU Jan 16, 2025

Choose a reason for hiding this comment

rust-log-analyzer commented Jan 20, 2025

Lower multi-segment const paths as `ConstArgKind::Path` #135186

Lower multi-segment const paths as `ConstArgKind::Path` #135186

camelid commented Jan 7, 2025 •

edited

Loading