Skip to content
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

Treat weak alias types more like ADTs when computing implied bounds #122340

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions compiler/rustc_hir_analysis/src/outlives/utils.rs
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@ pub(crate) fn insert_outlives_predicate<'tcx>(
) {
// If the `'a` region is bound within the field type itself, we
// don't want to propagate this constraint to the header.
if !is_free_region(outlived_region) {
if !is_early_bound_region(outlived_region) {
return;
}

Expand Down Expand Up @@ -132,7 +132,7 @@ pub(crate) fn insert_outlives_predicate<'tcx>(
}

GenericArgKind::Lifetime(r) => {
if !is_free_region(r) {
if !is_early_bound_region(r) {
return;
}
required_predicates.entry(ty::OutlivesPredicate(kind, outlived_region)).or_insert(span);
Expand All @@ -144,7 +144,7 @@ pub(crate) fn insert_outlives_predicate<'tcx>(
}
}

fn is_free_region(region: Region<'_>) -> bool {
fn is_early_bound_region(region: Region<'_>) -> bool {
Copy link
Member Author

@fmease fmease Mar 11, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just a drive-by renaming, cc #117908. I don't know why this used the term free (ReLateParam). It should've always been named is_early_bound_region or sth. in that vein.

// First, screen for regions that might appear in a type header.
match *region {
// These correspond to `T: 'a` relationships:
Expand Down
33 changes: 17 additions & 16 deletions compiler/rustc_infer/src/infer/outlives/components.rs
Original file line number Diff line number Diff line change
Expand Up @@ -138,7 +138,7 @@ fn compute_components<'tcx>(
// trait-ref. Therefore, if we see any higher-ranked regions,
// we simply fallback to the most restrictive rule, which
// requires that `Pi: 'a` for all `i`.
ty::Alias(_, alias_ty) => {
ty::Alias(ty::Projection | ty::Inherent | ty::Opaque, alias_ty) => {
if !alias_ty.has_escaping_bound_vars() {
// best case: no escaping regions, so push the
// projection and skip the subtree (thus generating no
Expand Down Expand Up @@ -171,21 +171,22 @@ fn compute_components<'tcx>(
// the type and then visits the types that are lexically
// contained within. (The comments refer to relevant rules
// from RFC1214.)
ty::Bool | // OutlivesScalar
ty::Char | // OutlivesScalar
ty::Int(..) | // OutlivesScalar
ty::Uint(..) | // OutlivesScalar
ty::Float(..) | // OutlivesScalar
ty::Never | // ...
ty::Adt(..) | // OutlivesNominalType
ty::Foreign(..) | // OutlivesNominalType
ty::Str | // OutlivesScalar (ish)
ty::Slice(..) | // ...
ty::RawPtr(..) | // ...
ty::Ref(..) | // OutlivesReference
ty::Tuple(..) | // ...
ty::FnPtr(_) | // OutlivesFunction (*)
ty::Dynamic(..) | // OutlivesObject, OutlivesFragment (*)
ty::Bool | // OutlivesScalar
ty::Char | // OutlivesScalar
ty::Int(..) | // OutlivesScalar
ty::Uint(..) | // OutlivesScalar
ty::Float(..) | // OutlivesScalar
ty::Never | // ...
ty::Adt(..) | // OutlivesNominalType
ty::Alias(ty::Weak, _) | // OutlivesNominalType (ish)
ty::Foreign(..) | // OutlivesNominalType
ty::Str | // OutlivesScalar (ish)
ty::Slice(..) | // ...
ty::RawPtr(..) | // ...
ty::Ref(..) | // OutlivesReference
ty::Tuple(..) | // ...
ty::FnPtr(_) | // OutlivesFunction (*)
ty::Dynamic(..) | // OutlivesObject, OutlivesFragment (*)
ty::Bound(..) |
ty::Error(_) => {
// (*) Function pointers and trait objects are both binders.
Expand Down
11 changes: 11 additions & 0 deletions tests/ui/lazy-type-alias/implied-outlives-bounds-1.print.stderr
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
error: rustc_outlives
--> $DIR/implied-outlives-bounds-1.rs:16:1
|
LL | struct Type<'a, K, V>(&'a mut Alias<K, V>);
| ^^^^^^^^^^^^^^^^^^^^^
|
= note: K: 'a
= note: V: 'a

error: aborting due to 1 previous error

20 changes: 20 additions & 0 deletions tests/ui/lazy-type-alias/implied-outlives-bounds-1.rs
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
// Check that we infer the outlives-predicates `K: 'a`, `V: 'a` for `Type`
// from the weak alias `Alias`.
// This mirrors the behavior of ADTs instead of other kinds of alias types
// like projections and opaque types.
// If we were to mirror the semantics of the latter, we would infer the
// outlives-predicate `Alias<K, V>: 'a` instead which is not what we want.

//@ revisions: default print
//@[default] check-pass

#![feature(lazy_type_alias)]
#![cfg_attr(print, feature(rustc_attrs))]
#![allow(incomplete_features)]

#[cfg_attr(print, rustc_outlives)]
struct Type<'a, K, V>(&'a mut Alias<K, V>); //[print]~ ERROR rustc_outlives

type Alias<K, V> = (K, V);

fn main() {}
2 changes: 0 additions & 2 deletions tests/ui/type-alias-impl-trait/implied_lifetime_wf_check3.rs
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,6 @@ mod test_lifetime_param {
fn defining(a: &str) -> Ty<'_> { a }
fn assert_static<'a: 'static>() {}
fn test<'a>() where Ty<'a>: 'static { assert_static::<'a>() }
//~^ ERROR: lifetime may not live long enough
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

👀

Copy link
Member Author

@fmease fmease Mar 11, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

WeakTy<'co>: 'static now implies 'co: 'static similar to ADTs. Is that fine / sound?

Copy link
Contributor

@oli-obk oli-obk Mar 19, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

it is, but we need tests that show that we actually enforce it, too. I don't know if we have any weak alias type tests for that, even if we have old-style type alias tests for it

}

mod test_higher_kinded_lifetime_param {
Expand All @@ -27,7 +26,6 @@ mod test_type_param {
fn defining<A>(s: A) -> Ty<A> { s }
fn assert_static<A: 'static>() {}
fn test<A>() where Ty<A>: 'static { assert_static::<A>() }
//~^ ERROR: parameter type `A` may not live long enough
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

👀

Copy link
Member Author

@fmease fmease Mar 11, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Similarly, WeakTy<Co>: 'static now implies Co: 'static similar to ADTs. Is that acceptable?

}

mod test_implied_from_fn_sig {
Expand Down
27 changes: 3 additions & 24 deletions tests/ui/type-alias-impl-trait/implied_lifetime_wf_check3.stderr
Original file line number Diff line number Diff line change
@@ -1,37 +1,16 @@
error: lifetime may not live long enough
--> $DIR/implied_lifetime_wf_check3.rs:7:43
|
LL | fn test<'a>() where Ty<'a>: 'static { assert_static::<'a>() }
| -- lifetime `'a` defined here ^^^^^^^^^^^^^^^^^^^ requires that `'a` must outlive `'static`

error: lifetime may not live long enough
--> $DIR/implied_lifetime_wf_check3.rs:15:46
--> $DIR/implied_lifetime_wf_check3.rs:14:46
|
LL | fn test<'a>() where for<'b> Ty<'b>: 'a { assert_static::<'a>() }
| -- lifetime `'a` defined here ^^^^^^^^^^^^^^^^^^^ requires that `'a` must outlive `'static`

error: lifetime may not live long enough
--> $DIR/implied_lifetime_wf_check3.rs:21:21
--> $DIR/implied_lifetime_wf_check3.rs:20:21
|
LL | fn test<'a>() { assert_static::<'a>() }
| -- ^^^^^^^^^^^^^^^^^^^ requires that `'a` must outlive `'static`
| |
| lifetime `'a` defined here

error[E0310]: the parameter type `A` may not live long enough
--> $DIR/implied_lifetime_wf_check3.rs:29:41
|
LL | fn test<A>() where Ty<A>: 'static { assert_static::<A>() }
| ^^^^^^^^^^^^^^^^^^
| |
| the parameter type `A` must be valid for the static lifetime...
| ...so that the type `A` will meet its required lifetime bounds
|
help: consider adding an explicit lifetime bound
|
LL | fn test<A: 'static>() where Ty<A>: 'static { assert_static::<A>() }
| +++++++++

error: aborting due to 4 previous errors
error: aborting due to 2 previous errors

For more information about this error, try `rustc --explain E0310`.
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,6 @@ mod test_type_param_static {
fn defining<A: 'static>(s: A) -> Ty<A> { s }
fn assert_static<A: 'static>() {}
fn test<A>() where Ty<A>: 'static { assert_static::<A>() }
//~^ ERROR: the parameter type `A` may not live long enough
}

fn main() {}
Original file line number Diff line number Diff line change
Expand Up @@ -12,20 +12,6 @@ help: consider adding an explicit lifetime bound
LL | type Ty<A: 'static> = impl Sized + 'static;
| +++++++++

error[E0310]: the parameter type `A` may not live long enough
--> $DIR/implied_lifetime_wf_check4_static.rs:8:41
|
LL | fn test<A>() where Ty<A>: 'static { assert_static::<A>() }
| ^^^^^^^^^^^^^^^^^^
| |
| the parameter type `A` must be valid for the static lifetime...
| ...so that the type `A` will meet its required lifetime bounds
|
help: consider adding an explicit lifetime bound
|
LL | fn test<A: 'static>() where Ty<A>: 'static { assert_static::<A>() }
| +++++++++

error: aborting due to 2 previous errors
error: aborting due to 1 previous error

For more information about this error, try `rustc --explain E0310`.
Loading