[WIP] mgca: Add ConstArg representation for const items #139558
Draft
+521
−386
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
rustbot
added
S-waiting-on-review
camelid
added
S-waiting-on-author
This comment has been minimized.
This comment has been minimized.
camelid
force-pushed
the
mgca-const-items
branch
from
fa42f86 to
6054bd5
Compare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
camelid
force-pushed
the
mgca-const-items
branch
from
ac73a4a to
4f6c9ab
Compare
camelid
commented
Apr 28, 2025
| let ty = this | ||
| .lower_ty(ty, ImplTraitContext::Disallowed(ImplTraitPosition::ConstTy)); | ||
| let body = | ||
| this.lower_const_item(span, body_id.unwrap(), expr.as_deref().unwrap()); |
There was a problem hiding this comment.
These new unwraps should be correct in theory since an error should've already been emitted, but they might ICE if the compiler didn't stop earlier. It might be a good idea to add ConstArgKind::Err and use delayed_span_bug.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
rustbot
added
the
WG-trait-system-refactor
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
camelid
added
A-const-generics
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
☔ The latest upstream changes (presumably #141343) made this pull request unmergeable. Please resolve the merge conflicts. |
BoxyUwU
reviewed
May 22, 2025
BoxyUwU
reviewed
May 22, 2025
Comment on lines
-217
to
-211
| DefKind::Const if tcx.generics_of(item_def_id).is_empty() => { | ||
| let instance = ty::Instance::new_raw(item_def_id.into(), ty::GenericArgs::empty()); | ||
| let cid = GlobalId { instance, promoted: None }; | ||
| let typing_env = ty::TypingEnv::fully_monomorphized(); | ||
| tcx.ensure_ok().eval_to_const_value_raw(typing_env.as_query_input(cid)); | ||
| } |
There was a problem hiding this comment.
We should properly handle this before landing 🤔 don't need to care too much rn while everything is still ICEing. I expect we'll just want to update fn check_associated_item and fn check_item in wfcheck.rs to handle const aliases like we do type aliases and then get rid of this codepath
BoxyUwU
reviewed
May 23, 2025
BoxyUwU
reviewed
May 23, 2025
Comment on lines
+233
to
+238
| /// Returns the const of the RHS of a const item. | ||
| query const_of_item(def_id: DefId) -> ty::EarlyBinder<'tcx, ty::Const<'tcx>> { | ||
| desc { |tcx| "computing the value for `{}`", tcx.def_path_str(def_id) } | ||
| cache_on_disk_if { def_id.is_local() } | ||
| separate_provide_extern | ||
| } |
There was a problem hiding this comment.
You'll need to do some more stuff to get this separate_provide_extern to actually work. You'll probably start getting ICEs in cross crate scenarios once you're past building core. See this commit for an example of adding a separate_provide_extern query: 996a185
BoxyUwU
reviewed
May 23, 2025
Comment on lines
+101
to
+103
| if let Ok(Some(Instance { def: ty::InstanceKind::Item(def_id), args })) = result | ||
| && matches!(tcx.def_kind(def_id), DefKind::Const | DefKind::AssocConst) | ||
| { |
There was a problem hiding this comment.
So it turns out that not all CTFE entry points actually call try_resolve, some of them just manually construct an instance for a const item. Not sure what to do about that long term 🤔
For now though you can go through all the entry points in rustc_middle/src/mir/interpret/queries and update everything to use Instance::expect_resolve instead of Instance::new_raw which should atleast solve any ICEs in the short term until we figure out how to properly handle this
|
@oli-obk I've assigned you to this PR alongside me because I'd definitely want you to review it before it lands since its so CTFE involved 🤔 I don't think it needs reviewing rn though, things are so up in the air and we're not bootstrapping yet :3 |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
BoxyUwU
reviewed
May 25, 2025
Comment on lines
+111
to
+115
| ty::AnonConstKind::MCG | ||
| if matches!(parent_def_kind, DefKind::Const | DefKind::AssocConst) => | ||
| { | ||
| Some(parent_did) | ||
| } |
There was a problem hiding this comment.
This will probably make the anon const in const FOO: [u8; ...] have generics too which is not desirable 🤔 Should look at the HIR instead to see if the parent hir node of the ConstArg node is a const item or not. Not sure whether it makes sense to add a new variant to ty::AnonConstKind or not, I guess it depends on whether we wind up special casing them elsewhere (e.g. HIR ty lowering).
Actually that reminds me- we should have a test about const ASSOC: usize = (N);. It's possible that will get lowered to an anon const instead of a path (expected) and then ICE in ty lowering due to being a bare usage of a generic parameter not lowered to a path. Properly fixing this might just require changing ty lowering to not do anything fancy for anon consts used as the rhs of const items, which probably motivates adding a ty::AnonConstKind::ConstBody
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
The job Click to see the possible cause of the failure (guessed by this bot) |
|
☔ The latest upstream changes (presumably #141605) made this pull request unmergeable. Please resolve the merge conflicts. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
r? @BoxyUwU