From 558a41e3c95338590b172719e6dd25215d4e8450 Mon Sep 17 00:00:00 2001
From: Tobiasz Laskowski <tobil4sk@outlook.com>
Date: Tue, 25 Aug 2026 02:40:01 +0100
Subject: [PATCH 1/4] Fix String::hash for utf16 strings

---
 include/hxString.h | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/include/hxString.h b/include/hxString.h
index a140167eb..5fd402561 100644
--- a/include/hxString.h
+++ b/include/hxString.h
@@ -228,8 +228,8 @@ class HXCPP_EXTERN_CLASS_ATTRIBUTES String
          #ifdef HXCPP_PARANOID
          unsigned int result = calcHash();
 
-         unsigned int have = (((unsigned int *)__s)[-1] & HX_GC_CONST_ALLOC_BIT) ?
-                ((unsigned int *)__s)[-2] :  *((unsigned int *)(__s+length+1) );
+         unsigned int have = (((unsigned int *)__s)[-1] & HX_GC_CONST_ALLOC_BIT) ? ((unsigned int *)__s)[-2] :
+                             isUTF16Encoded() ? *((unsigned int *)(__w+length+1)) : *((unsigned int *)(__s+length+1));
 
          if ( have != result )
          {
@@ -247,6 +247,13 @@ class HXCPP_EXTERN_CLASS_ATTRIBUTES String
             return  ((unsigned int *)__s)[-2];
             #endif
          }
+         if (isUTF16Encoded()) {
+            #ifdef __EMSCRIPTEN__
+            return *((emscripten_align1_int *)(__w+length+1));
+            #else
+            return *((unsigned int *)(__w+length+1));
+            #endif
+         }
         #ifdef __EMSCRIPTEN__
            return *((emscripten_align1_int *)(__s+length+1) );
         #else

From 98865cc897e1b0b17ad99c58f6251caed92ec957 Mon Sep 17 00:00:00 2001
From: Tobiasz Laskowski <tobil4sk@outlook.com>
Date: Tue, 25 Aug 2026 03:22:16 +0100
Subject: [PATCH 2/4] [tests] Verify unicode string hashing

---
 test/haxe/TestStringHash.hx | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/test/haxe/TestStringHash.hx b/test/haxe/TestStringHash.hx
index 2cdaa3635..793a6ab4f 100644
--- a/test/haxe/TestStringHash.hx
+++ b/test/haxe/TestStringHash.hx
@@ -80,4 +80,21 @@ class TestStringHash extends Test
       Assert.pass();
    }
 
+   public function testUnicode()
+   {
+		for (literal in ["á", "👍"]) {
+			var fromBytes = haxe.io.Bytes.ofString(literal).toString();
+
+			Assert.equals(literal, fromBytes);
+
+			var m = new Map<String, Int>();
+			m[literal] = 1;
+
+			Assert.isTrue(m.exists(literal));
+			Assert.isTrue(m.exists(fromBytes));
+
+			Assert.equals(1, m[literal]);
+			Assert.equals(1, m[fromBytes]);
+      }
+   }
 }

From c61a4cca8fbdf0997af668219836d04c95c55ab4 Mon Sep 17 00:00:00 2001
From: Tobiasz Laskowski <tobil4sk@outlook.com>
Date: Tue, 25 Aug 2026 03:34:38 +0100
Subject: [PATCH 3/4] Fix String::calcHash with non-BMP characters

Outside the BMP, characters take two utf16 code units. The old hashing
was incorrect as it would only ever read a single unit at a time, and
therefore calculate the wrong hash for a string containing a non-BMP
character.

The Char16Advance call handles this, ensuring that surrogate pairs are
read as a single code point.
---
 src/String.cpp | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/src/String.cpp b/src/String.cpp
index 4d84b77b7..dd0cbca22 100644
--- a/src/String.cpp
+++ b/src/String.cpp
@@ -932,9 +932,10 @@ unsigned int String::calcSubHash(int start, int inLen) const
    if (isUTF16Encoded())
    {
       const char16_t *w = __w + start;
-      for(int i=0;i<inLen;i++)
+      const char16_t *end = w + inLen;
+      while (w < end)
       {
-         int c = w[i];
+         int c = Char16Advance(w, false);
          if( c <= 0x7F )
          {
             ADD_HASH(c);
@@ -977,9 +978,11 @@ unsigned int String::calcHash() const
    #ifdef HX_SMART_STRINGS
    if (isUTF16Encoded())
    {
-      for(int i=0;i<length;i++)
+      const char16_t *w = __w;
+      const char16_t *end = w + length;
+      while (w < end)
       {
-         int c = __w[i];
+         int c = Char16Advance(w, false);
          if( c <= 0x7F )
          {
             ADD_HASH(c);

From 769fe85e6768d124523bd6bff258454e49971218 Mon Sep 17 00:00:00 2001
From: Tobiasz Laskowski <tobil4sk@outlook.com>
Date: Tue, 25 Aug 2026 03:35:36 +0100
Subject: [PATCH 4/4] [tests] Verify non-BMP string hashing

---
 test/haxe/TestStringHash.hx | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/test/haxe/TestStringHash.hx b/test/haxe/TestStringHash.hx
index 793a6ab4f..19a2077fc 100644
--- a/test/haxe/TestStringHash.hx
+++ b/test/haxe/TestStringHash.hx
@@ -83,17 +83,22 @@ class TestStringHash extends Test
    public function testUnicode()
    {
 		for (literal in ["á", "👍"]) {
+			var fromJson:String = haxe.Json.parse('\"$literal\"');
 			var fromBytes = haxe.io.Bytes.ofString(literal).toString();
 
+			Assert.equals(literal, fromJson);
+			Assert.equals(fromJson, fromBytes);
 			Assert.equals(literal, fromBytes);
 
 			var m = new Map<String, Int>();
 			m[literal] = 1;
 
 			Assert.isTrue(m.exists(literal));
+			Assert.isTrue(m.exists(fromJson));
 			Assert.isTrue(m.exists(fromBytes));
 
 			Assert.equals(1, m[literal]);
+			Assert.equals(1, m[fromJson]);
 			Assert.equals(1, m[fromBytes]);
       }
    }