Rollup merge of #126980 - Borgerr:fix-extendfromslice-check, r=workingjubilee

set self.is_known_utf8 to false in extend_from_slice

try-job: x86_64-msvc

closes #126977
Related to #126885, #126333, and [this conversation](<aa46a3368e (r143539097)>)

library/std/src/sys_common/wtf8.rs

@@ -480,7 +480,7 @@ pub fn from_box(boxed: Box<Wtf8>) -> 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;
     }
 }
 

library/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);
+}