mirror of
https://github.com/rust-lang/rust.git
synced 2024-11-01 06:51:58 +00:00
perf: only check for rustc_trivial_field_reads attribute on traits
The checks removed here caused a small perf regression: https://github.com/rust-lang/rust/pull/88824#issuecomment-932664761 Since the attribute is currently only applied to traits, I don't think it's worth keeping the additional checks for now. If/when we decide to apply the attribute somewhere else, we can (partially) revert this and evaluate if the perf impact is acceptable.
This commit is contained in:
parent
c02371c442
commit
eb4ba58572
@ -254,31 +254,11 @@ impl<'tcx> MarkSymbolVisitor<'tcx> {
|
||||
|
||||
let has_attr = |def_id| self.tcx.has_attr(def_id, sym::rustc_trivial_field_reads);
|
||||
|
||||
if has_attr(def_id) {
|
||||
return true;
|
||||
}
|
||||
|
||||
if let Some(impl_of) = self.tcx.impl_of_method(def_id) {
|
||||
if has_attr(impl_of) {
|
||||
return true;
|
||||
}
|
||||
|
||||
if let Some(trait_of) = self.tcx.trait_id_of_impl(impl_of) {
|
||||
if has_attr(trait_of) {
|
||||
return true;
|
||||
}
|
||||
|
||||
if let Some(method_ident) = self.tcx.opt_item_name(def_id) {
|
||||
if let Some(trait_method) = self
|
||||
.tcx
|
||||
.associated_items(trait_of)
|
||||
.find_by_name_and_kind(self.tcx, method_ident, ty::AssocKind::Fn, trait_of)
|
||||
{
|
||||
if has_attr(trait_method.def_id) {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
} else if let Some(trait_of) = self.tcx.trait_of_item(def_id) {
|
||||
if has_attr(trait_of) {
|
||||
@ -291,9 +271,7 @@ impl<'tcx> MarkSymbolVisitor<'tcx> {
|
||||
|
||||
fn visit_node(&mut self, node: Node<'tcx>) {
|
||||
if let Some(item_def_id) = match node {
|
||||
Node::Item(hir::Item { def_id, .. })
|
||||
| Node::ForeignItem(hir::ForeignItem { def_id, .. })
|
||||
| Node::TraitItem(hir::TraitItem { def_id, .. })
|
||||
Node::TraitItem(hir::TraitItem { def_id, .. })
|
||||
| Node::ImplItem(hir::ImplItem { def_id, .. }) => Some(def_id.to_def_id()),
|
||||
_ => None,
|
||||
} {
|
||||
|
Loading…
Reference in New Issue
Block a user