From 57c2de81f2b53dc7cb93f8ba0d3e24f80cd61425 Mon Sep 17 00:00:00 2001
From: ash <97464181+Borgerr@users.noreply.github.com>
Date: Tue, 25 Jun 2024 23:58:43 -0600
Subject: [PATCH 1/2] set self.is_known_utf8 to false in extend_from_slice

---
 std/src/sys_common/wtf8.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/std/src/sys_common/wtf8.rs b/std/src/sys_common/wtf8.rs
index 708f62f476e73..117a3e23044ea 100644
--- a/std/src/sys_common/wtf8.rs
+++ b/std/src/sys_common/wtf8.rs
@@ -480,7 +480,7 @@ impl Wtf8Buf {
     #[inline]
     pub(crate) fn extend_from_slice(&mut self, other: &[u8]) {
         self.bytes.extend_from_slice(other);
-        self.is_known_utf8 = self.is_known_utf8 || self.next_surrogate(0).is_none();
+        self.is_known_utf8 = false;
     }
 }
 

From 4788a93eee24a3e90becec001fd921e5565330db Mon Sep 17 00:00:00 2001
From: Jubilee Young <workingjubilee@gmail.com>
Date: Tue, 25 Jun 2024 22:29:37 -0700
Subject: [PATCH 2/2] std: test a variety of ways to extend a Wtf8Buf

---
 std/src/sys_common/wtf8/tests.rs | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/std/src/sys_common/wtf8/tests.rs b/std/src/sys_common/wtf8/tests.rs
index 6a1cc41a8fb04..b57c99a8452a1 100644
--- a/std/src/sys_common/wtf8/tests.rs
+++ b/std/src/sys_common/wtf8/tests.rs
@@ -725,3 +725,27 @@ fn wtf8_utf8_boundary_between_surrogates() {
     string.push(CodePoint::from_u32(0xD800).unwrap());
     check_utf8_boundary(&string, 3);
 }
+
+#[test]
+fn wobbled_wtf8_plus_bytes_isnt_utf8() {
+    let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() };
+    assert!(!string.is_known_utf8);
+    string.extend_from_slice(b"some utf-8");
+    assert!(!string.is_known_utf8);
+}
+
+#[test]
+fn wobbled_wtf8_plus_str_isnt_utf8() {
+    let mut string: Wtf8Buf = unsafe { Wtf8::from_bytes_unchecked(b"\xED\xA0\x80").to_owned() };
+    assert!(!string.is_known_utf8);
+    string.push_str("some utf-8");
+    assert!(!string.is_known_utf8);
+}
+
+#[test]
+fn unwobbly_wtf8_plus_utf8_is_utf8() {
+    let mut string: Wtf8Buf = Wtf8Buf::from_str("hello world");
+    assert!(string.is_known_utf8);
+    string.push_str("some utf-8");
+    assert!(string.is_known_utf8);
+}