use a must_hold variant for checking PartialEq

This commit is contained in:
Ralf Jung 2023-09-24 16:37:19 +02:00
parent c3ed0c454e
commit b589976606

View File

@ -179,7 +179,7 @@ impl<'tcx> ConstToPat<'tcx> {
}
if let Some(non_sm_ty) = structural {
if !self.type_may_have_partial_eq_impl(cv.ty()) {
if !self.type_has_partial_eq_impl(cv.ty()) {
if let ty::Adt(def, ..) = non_sm_ty.kind() {
if def.is_union() {
let err = UnionPattern { span: self.span };
@ -244,7 +244,7 @@ impl<'tcx> ConstToPat<'tcx> {
// Always check for `PartialEq`, even if we emitted other lints. (But not if there were
// any errors.) This ensures it shows up in cargo's future-compat reports as well.
if !self.type_may_have_partial_eq_impl(cv.ty()) {
if !self.type_has_partial_eq_impl(cv.ty()) {
self.tcx().emit_spanned_lint(
lint::builtin::MATCH_WITHOUT_PARTIAL_EQ,
self.id,
@ -258,7 +258,7 @@ impl<'tcx> ConstToPat<'tcx> {
}
#[instrument(level = "trace", skip(self), ret)]
fn type_may_have_partial_eq_impl(&self, ty: Ty<'tcx>) -> bool {
fn type_has_partial_eq_impl(&self, ty: Ty<'tcx>) -> bool {
// double-check there even *is* a semantic `PartialEq` to dispatch to.
//
// (If there isn't, then we can safely issue a hard
@ -273,8 +273,13 @@ impl<'tcx> ConstToPat<'tcx> {
ty::TraitRef::new(self.tcx(), partial_eq_trait_id, [ty, ty]),
);
// FIXME: should this call a `predicate_must_hold` variant instead?
self.infcx.predicate_may_hold(&partial_eq_obligation)
// This *could* accept a type that isn't actually `PartialEq`, because region bounds get
// ignored. However that should be pretty much impossible since consts that do not depend on
// generics can only mention the `'static` lifetime, and how would one have a type that's
// `PartialEq` for some lifetime but *not* for `'static`? If this ever becomes a problem
// we'll need to leave some sort of trace of this requirement in the MIR so that borrowck
// can ensure that the type really implements `PartialEq`.
self.infcx.predicate_must_hold_modulo_regions(&partial_eq_obligation)
}
fn field_pats(