From e732d1f273297da21126ee1f0b0060a513685557 Mon Sep 17 00:00:00 2001
From: Aapo Alasuutari <aapo.alasuutari@gmail.com>
Date: Sat, 7 Nov 2026 14:05:52 +0200
Subject: [PATCH 1/2] fix(small_string): utf8_index

---
 small_string/lib.rs | 30 +++++++++++++++++++++++++-----
 1 file changed, 25 insertions(+), 5 deletions(-)

diff --git a/small_string/lib.rs b/small_string/lib.rs
index bc6674570..84cbfe88b 100644
--- a/small_string/lib.rs
+++ b/small_string/lib.rs
@@ -207,26 +207,35 @@ impl SmallString {
     pub fn utf8_index(&self, utf16_idx: usize) -> Option<usize> {
         if self.is_ascii() {
             return Some(utf16_idx);
+        } else if utf16_idx >= self.utf16_len() {
+            return None;
         }
         let mut current_utf16_index = 0;
+        let mut current_utf8_index = 0;
         let mut scratch = [0u16; 2];
-        for (idx, ch) in self.as_wtf8().code_points().enumerate() {
+        for ch in self.as_wtf8().code_points() {
             match current_utf16_index.cmp(&utf16_idx) {
-                Ordering::Equal => return Some(idx),
+                Ordering::Equal => return Some(current_utf8_index),
                 Ordering::Greater => return None,
                 Ordering::Less => {
                     current_utf16_index += ch
                         .to_char()
                         .map(|ch| ch.encode_utf16(&mut scratch).len())
-                        .unwrap_or(1)
+                        .unwrap_or(1);
+                    // Lone surrogates take 3 bytes in WTF-8.
+                    current_utf8_index += match ch.to_u32() {
+                        0..=0x7F => 1,
+                        0x80..=0x7FF => 2,
+                        0x800..=0xFFFF => 3,
+                        _ => 4,
+                    };
                 }
             }
         }
         if current_utf16_index > utf16_idx {
             return None;
         }
-        debug_assert_eq!(utf16_idx, current_utf16_index);
-        Some(self.len())
+        Some(current_utf8_index)
     }
 
     /// Get the corresponding WTF-16 code unit index for a given UTF-8 code
@@ -615,3 +624,14 @@ fn str_conversion() {
     assert!(SmallString::try_from(too_large_unicode).is_err());
     assert!(SmallString::try_from(Wtf8::from_str(too_large_unicode)).is_err());
 }
+
+#[test]
+fn utf8_indexing() {
+    let unicode = "あ<";
+    let str = SmallString::try_from(unicode).unwrap();
+    assert_eq!(str.len(), 4);
+    assert_eq!(str, unicode);
+    assert_eq!(str.utf8_index(0), Some(0));
+    assert_eq!(str.utf8_index(1), Some(3));
+    assert_eq!(str.utf8_index(2), None);
+}

From e6086f3733249ce538c5d2cdb6626eaf49b7135a Mon Sep 17 00:00:00 2001
From: Aapo Alasuutari <aapo.alasuutari@gmail.com>
Date: Sat, 7 Nov 2026 14:14:40 +0200
Subject: [PATCH 2/2] fix(ecmascript): mapping of zero index in heap strings

---
 nova_vm/src/ecmascript/types/language/string/data.rs | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/nova_vm/src/ecmascript/types/language/string/data.rs b/nova_vm/src/ecmascript/types/language/string/data.rs
index 2d7f5fea0..5f45119f0 100644
--- a/nova_vm/src/ecmascript/types/language/string/data.rs
+++ b/nova_vm/src/ecmascript/types/language/string/data.rs
@@ -163,8 +163,13 @@ impl StringRecord {
                             // We got a None; that means we're looking at a latter
                             // surrogate here.
                             // SAFETY: idx is not 0.
-                            let idx = mapping[unsafe { idx.unchecked_sub(1) }].unwrap();
-                            (idx.into(), true)
+                            let idx = unsafe { idx.unchecked_sub(1) };
+                            let idx = if idx == 0 {
+                                0
+                            } else {
+                                mapping[idx].unwrap().into()
+                            };
+                            (idx, true)
                         }
                     }
                 }