Update (doc) comments

Several (doc) comments were super outdated or didn't provide enough context.

Some doc comments shoved everything in a single paragraph without respecting
the fact that the first paragraph should be a single sentence because rustdoc
treats these as item descriptions / synopses on module pages.

compiler/rustc_trait_selection/src/solve/normalizes_to/inherent.rs

@@ -1,5 +1,5 @@
 //! Computes a normalizes-to (projection) goal for inherent associated types,
-//! `#![feature(inherent_associated_type)]`. Since astconv already determines
+//! `#![feature(inherent_associated_type)]`. Since HIR ty lowering already determines
 //! which impl the IAT is being projected from, we just:
 //! 1. instantiate generic parameters,
 //! 2. equate the self type, and

compiler/rustc_trait_selection/src/traits/object_safety.rs

@@ -32,8 +32,9 @@ use std::ops::ControlFlow;
 
 pub use crate::traits::{MethodViolationCode, ObjectSafetyViolation};
 
-/// Returns the object safety violations that affect
-/// astconv -- currently, `Self` in supertraits. This is needed
+/// Returns the object safety violations that affect HIR ty lowering.
+///
+/// Currently that is `Self` in supertraits. This is needed
 /// because `object_safety_violations` can't be used during
 /// type collection.
 pub fn hir_ty_lowering_object_safety_violations(

compiler/rustc_trait_selection/src/traits/select/confirmation.rs

@@ -1264,7 +1264,7 @@ impl<'cx, 'tcx> SelectionContext<'cx, 'tcx> {
 
                 // Extract `TailField<T>` and `TailField<U>` from `Struct<T>` and `Struct<U>`,
                 // normalizing in the process, since `type_of` returns something directly from
-                // astconv (which means it's un-normalized).
+                // HIR ty lowering (which means it's un-normalized).
                 let source_tail = normalize_with_depth_to(
                     self,
                     obligation.param_env,

compiler/rustc_trait_selection/src/traits/specialize/specialization_graph.rs

@@ -396,7 +396,7 @@ pub(crate) fn assoc_def(
         // associated type. Normally this situation
         // could only arise through a compiler bug --
         // if the user wrote a bad item name, it
-        // should have failed in astconv.
+        // should have failed during HIR ty lowering.
         bug!(
             "No associated type `{}` for {}",
             tcx.item_name(assoc_def_id),