rustdoc: fix ICE when a grapheme cluster joins a Prepend-class character to _ or :

src/librustdoc/html/escape.rs

@@ -79,9 +79,12 @@ impl fmt::Display for EscapeBodyTextWithWbr<'_> {
             } else if (s.contains(':') && !next_is_colon())
                 || (s.contains('_') && !next_is_underscore())
             {
-                EscapeBodyText(&text[last..i + 1]).fmt(fmt)?;
+                // `i + s.len()`, not `i + 1`: a cluster containing `_` or `:` can still be
+                // several bytes long, because UAX#29 GB9b joins a `Prepend` character to the
+                // one that follows it.
+                EscapeBodyText(&text[last..i + s.len()]).fmt(fmt)?;
                 fmt.write_str("<wbr>")?;
-                last = i + 1;
+                last = i + s.len();
             }
         }
         if last < text.len() {

src/librustdoc/html/escape/tests.rs

@@ -40,6 +40,10 @@ fn escape_body_text_with_wbr() {
     assert_eq!(&E("ṼẽçÑñéå").to_string(), "Ṽẽç<wbr>Ññéå");
     assert_eq!(&E("V\u{0300}e\u{0300}c\u{0300}D\u{0300}e\u{0300}q\u{0300}u\u{0300}e\u{0300}u\u{0300}e\u{0300}").to_string(), "V\u{0300}e\u{0300}c\u{0300}<wbr>D\u{0300}e\u{0300}q\u{0300}u\u{0300}e\u{0300}u\u{0300}e\u{0300}");
     assert_eq!(&E("LPFNACCESSIBLEOBJECTFROMWINDOW").to_string(), "LPFNACCESSIBLEOBJECTFROMWINDOW");
+    // clusters where the `_` or `:` is not the first byte (UAX#29 GB9b `Prepend`)
+    assert_eq!(&E("abc\u{0D4E}_defgh").to_string(), "abc\u{0D4E}_<wbr>defgh");
+    assert_eq!(&E("abc\u{0605}:defgh").to_string(), "abc\u{0605}:<wbr>defgh");
+    assert_eq!(&E("first_\u{0300}second").to_string(), "first_\u{0300}<wbr>second");
 }
 // property test
 #[test]

tests/rustdoc-html/item-name-grapheme-cluster-wbr.rs

@@ -0,0 +1,8 @@
+// Regression test for #160231: `<wbr>` insertion sliced mid-character on an item name whose
+// grapheme cluster starts with a `Prepend` character and contains `_`.
+
+#![crate_name = "foo"]
+#![allow(mixed_script_confusables, non_camel_case_types)]
+
+//@ hasraw foo/index.html 'abcൎ_<wbr>defgh'
+pub struct abcൎ_defgh;