Fix Encoding regression (#111320)

* Fix Encoding regression * Feedback addressing * Fix the test
dotnet · Jan 14, 2025 · 30834e4 · 30834e4
1 parent a05d34b
commit 30834e4
Show file tree

Hide file tree

Showing 2 changed files with 24 additions and 1 deletion.
diff --git a/src/libraries/System.Text.Encoding.CodePages/src/System/Text/EncodingCharBuffer.cs b/src/libraries/System.Text.Encoding.CodePages/src/System/Text/EncodingCharBuffer.cs
@@ -68,9 +68,14 @@ internal unsafe bool AddChar(char ch)
             return AddChar(ch, 1);
         }
 
-
         internal unsafe bool AddChar(char ch1, char ch2, int numBytes)
         {
+            if (_chars is null)
+            {
+                _charCountResult += 2;
+                return true;
+            }
+
             // Need room for 2 chars
             if (_charEnd - _chars < 2)
             {

diff --git a/src/libraries/System.Text.Encoding.CodePages/tests/EncodingCodePages.cs b/src/libraries/System.Text.Encoding.CodePages/tests/EncodingCodePages.cs
@@ -519,6 +519,7 @@ public static void TestDefaultEncodings()
             Assert.Contains(mappedEncoding, CrossplatformDefaultEncodings().Union(CodePageInfo().Select(i => Map((int)i[0], (string)i[1]))));
 
             TestRegister1252();
+            TestMultiBytesEncodingsSupportSurrogate();
         }
 
         private static void ValidateDefaultEncodings()
@@ -639,6 +640,23 @@ public static void TestEncodingDisplayNames(int codePage, string webName, string
             Assert.All(name, c => Assert.True(c >= ' ' && c < '~' + 1, "Name: " + name + " contains character: " + c));
         }
 
+        private static void TestMultiBytesEncodingsSupportSurrogate()
+        {
+            Encoding.RegisterProvider(CodePagesEncodingProvider.Instance);
+
+            Encoding encoding = Encoding.GetEncoding("GB18030");
+            Assert.NotNull(encoding);
+
+            string surrogatePair = "\uD840\uDE13"; // Surrogate Pair codepoint '𠈓' \U00020213
+            byte[] expectedBytes = new byte[] { 0x95, 0x32, 0xB7, 0x37 };
+
+            Assert.Equal(expectedBytes, encoding.GetBytes(surrogatePair));
+            Assert.Equal(expectedBytes.Length, encoding.GetByteCount(surrogatePair));
+
+            Assert.Equal(surrogatePair, encoding.GetString(expectedBytes));
+            Assert.Equal(surrogatePair.Length, encoding.GetCharCount(expectedBytes));
+        }
+
         // This test is run as part of the default mappings test, since it modifies global state which that test
         // depends on.
         private static void TestRegister1252()