-
Notifications
You must be signed in to change notification settings - Fork 12.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
handle ConstValue::ByRef in relate #71018
Conversation
= note: expected struct `Const<ByRef { alloc: Allocation { bytes: [3, 0, 0, 0, 0, 0, 0, 0], relocations: Relocations(SortedMap { data: [] }), undef_mask: UndefMask { blocks: [255], len: Size { raw: 8 } }, size: Size { raw: 8 }, align: Align { pow2: 3 }, mutability: Not, extra: () }, offset: Size { raw: 0 } }: [usize; 1]>` | ||
found struct `Const<ByRef { alloc: Allocation { bytes: [4, 0, 0, 0, 0, 0, 0, 0], relocations: Relocations(SortedMap { data: [] }), undef_mask: UndefMask { blocks: [255], len: Size { raw: 8 } }, size: Size { raw: 8 }, align: Align { pow2: 3 }, mutability: Not, extra: () }, offset: Size { raw: 0 } }: [usize; 1]>` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I have a WIP branch where I make this print the actual values, I think we should wait for that, so that we can actually see it in action.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
🤷♂️ don't think that waiting is worth it if this is otherwise ready.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's not ready, because we can't read the error message.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We already emit this error message rn. This PR only prevents correct programs from failing with it.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That branch I mentioned is now a PR: #71232.
src/librustc_middle/ty/relate.rs
Outdated
let a_destructured = tcx.destructure_const(param_env.and(a)); | ||
let b_destructured = tcx.destructure_const(param_env.and(b)); | ||
|
||
// Both the variant and each field have to be equal. | ||
if a_destructured.variant == b_destructured.variant { | ||
for (a_field, b_field) in | ||
a_destructured.fields.iter().zip(b_destructured.fields.iter()) | ||
{ | ||
relation.consts(a_field, b_field)?; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@oli-obk @matthewjasper What do you think of this use of destructure_const
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
destructure_const
panics if the type isn't an array, adt or tuple, so it looks incorrect to not have a check on the type.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it is. breaks with #63322
currently working on this
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It should allow closures as well, IMO, although I'm not sure closures should be considered structurally matcheable...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should be fixed, don't know if there are other types which have ConstValue::ByRef
and should work rn.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh yeah you can't have closures in a const
type, so you can mark closures as not structurally matcheable.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Wait, hang on, it can't be ty::Closure
, maybe you were still talking about dyn Trait
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
both examples in #63322 end up relying on &dyn Trait
afaict
56c3ec0
to
cb4ffda
Compare
cb4ffda
to
c1757a8
Compare
src/librustc_middle/ty/relate.rs
Outdated
// FIXME(const_generics): There are probably some `TyKind`s | ||
// which should be handled here. | ||
_ => Err(TypeError::ConstMismatch(expected_found(relation, &a, &b))), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I can't think of any, at least out of the types we allow. You should use at least delay_span_bug
, because returning Err
indicates that they're definitely not equal (which may allow e.g. overlapping impl
s).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
👌 should relate_consts
only be called for types which are structural match?
Is structural match implemented for &
?
i.e. should this work
#![feature(const_generics)]
struct A<const N: &'static ()>;
fn main() {
let _ = A::<{&()}>;
}
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, it should, but it's blocked on one of @oli-obk's PRs.
Blocking on #71018 (comment). |
c1757a8
to
ebacb34
Compare
ty/print: pretty-print constant aggregates (arrays, tuples and ADTs). Oddly enough, we don't have any UI tests showing this off in types, only `mir-opt` tests. However, the pretty form should show up in the test output diff of rust-lang#71018, if this PR is merged first. <hr/> Examples of before/after: |`Option<bool>`| |:-:| |`{transmute(0x01): std::option::Option<bool>}`| | :sparkles: ↓↓↓ :sparkles: | |`std::option::Option::<bool>::Some(true)`| | `RawVec<u32>` | |:-:| | `ByRef { alloc: Allocation { bytes: [4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], relocations: Relocations(SortedMap { data: [] }), undef_mask: UndefMask { blocks: [65535], len: Size { raw: 16 } }, size: Size { raw: 16 }, align: Align { pow2: 3 }, mutability: Not, extra: () }, offset: Size { raw: 0 } }: alloc::raw_vec::RawVec::<u32>`| | :sparkles: ↓↓↓ :sparkles: | |`alloc::raw_vec::RawVec::<u32> { ptr: std::ptr::Unique::<u32> { pointer: {0x4 as *const u32}, _marker: std::marker::PhantomData::<u32> }, cap: 0usize, alloc: std::alloc::Global }`| <hr/> This PR is a prerequisite for rust-lang#61486, *sort of*, in that we need to be able to pretty-print values in order to even consider how we might mangle them. We still don't have pretty-printing for constants of reference types, @oli-obk has the necessary support logic in a PR but I didn't want to interfere with that. <hr/> Each commit should be reviewed separately, as I've fixed a couple deficiencies along the way. r? @oli-obk cc @rust-lang/wg-mir-opt @varkor @yodaldevoid
I think this is unblocked now since the fore-mentioned PR has been merged, maybe this will need a rebase? |
ebacb34
to
8b0c6c4
Compare
Your PR failed (pretty log, raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem. Click to expand the log.
I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
8b0c6c4
to
b026d05
Compare
Should be ready for merge ✨ |
--> $DIR/different_byref.rs:8:9 | ||
| | ||
LL | x = Const::<{ [4] }> {}; | ||
| ^^^^^^^^^^^^^^^^^^^ expected `3usize`, found `4usize` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Only mildly related to this PR
@rust-lang/wg-const-eval we could consider collecting all spans that lead to a value during const eval and thus allow each byte in the final constant to have its own spans for diagnostics. Then we can report a sort of backtrace for values. Though I fear that will cause quite some overhead, even if we only need to do it in the current crate and not serialize it to metadata (so ClearCrossCrate
).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That would be more advanced than what we do for type aliases, heh.
☔ The latest upstream changes (presumably #71707) made this pull request unmergeable. Please resolve the merge conflicts. |
b026d05
to
a08bccb
Compare
@bors r+ |
📌 Commit a08bccb has been approved by |
handle ConstValue::ByRef in relate fixes rust-lang#68615 r? @eddyb
handle ConstValue::ByRef in relate fixes rust-lang#68615 r? @eddyb
⌛ Testing commit a08bccb with merge b6e7b292fff458a6728a49e39a114a5b00678970... |
@bors retry yield |
Rollup of 5 pull requests Successful merges: - rust-lang#71018 (handle ConstValue::ByRef in relate) - rust-lang#71758 (Remove leftover chalk types) - rust-lang#71760 (Document unsafety for `*const T` and `*mut T`) - rust-lang#71761 (doc: reference does not exist, probably a typo) - rust-lang#71762 (doc: this resulted in a link pointing to a non-existent target) Failed merges: - rust-lang#71726 (Suggest deref when coercing `ty::Ref` to `ty::RawPtr` with arbitrary mutability) r? @ghost
fixes #68615
r? @eddyb