Support UTF-8 surrogates for UTF-16 and 32.

aaubry · Aug 28, 2023 · e8d7113 · e8d7113
1 parent a6845eb
commit e8d7113
Show file tree

Hide file tree

Showing 3 changed files with 73 additions and 17 deletions.
diff --git a/YamlDotNet.Benchmark/Program.cs b/YamlDotNet.Benchmark/Program.cs
@@ -1,11 +1,4 @@
-using System.Globalization;
-using BenchmarkDotNet.Running;
+using BenchmarkDotNet.Running;
 using YamlDotNet.Benchmark;
-using YamlDotNet.Core;
-using YamlDotNet.Core.Events;
-using YamlDotNet.Serialization;
-using YamlDotNet.Serialization.NamingConventions;
 
-var dateTimeOffset = new DateTimeOffset(new DateTime(2017, 1, 2, 3, 4, 5), new TimeSpan(-6, 0, 0));
-Console.WriteLine(dateTimeOffset.ToString("MM/dd/yyyy HH:mm:ss zzz", CultureInfo.InvariantCulture));
-Console.WriteLine(dateTimeOffset.ToString("O", CultureInfo.InvariantCulture));
+BenchmarkSwitcher.FromAssembly(typeof(YamlStreamBenchmark).Assembly).Run(args);
diff --git a/YamlDotNet.Test/Core/ScannerTests.cs b/YamlDotNet.Test/Core/ScannerTests.cs
@@ -1,4 +1,4 @@
-// This file is part of YamlDotNet - A .NET library for YAML.
+// This file is part of YamlDotNet - A .NET library for YAML.
 // Copyright (c) Antoine Aubry and contributors
 //
 // Permission is hereby granted, free of charge, to any person obtaining a copy of
@@ -530,6 +530,20 @@ public void Keys_can_start_with_colons_after_double_quoted_values_in_nested_bloc
                 StreamEnd);
         }
 
+        [Fact]
+        public void Utf16StringsAsUtf8SurrogatesWorkCorrectly()
+        {
+            AssertSequenceOfTokensFrom(Yaml.ScannerForText("Test: \"\\uD83D\\uDC4D\""),
+                StreamStart,
+                BlockMappingStart,
+                Key,
+                PlainScalar("Test"),
+                Value,
+                DoubleQuotedScalar("\uD83D\uDC4D"), // guaranteed thumbs up emoticon that will work in Windows Terminal since it pukes on displaying it.
+                BlockEnd,
+                StreamEnd);
+        }
+
         private void AssertPartialSequenceOfTokensFrom(Scanner scanner, params Token[] tokens)
         {
             var tokenNumber = 1;

diff --git a/YamlDotNet/Core/Scanner.cs b/YamlDotNet/Core/Scanner.cs
@@ -1937,19 +1937,68 @@ private Scalar ScanFlowScalar(bool isSingleQuoted)
 
                             // Check the value and write the character.
 
-                            if ((character >= 0xD800 && character <= 0xDFFF) || character > 0x10FFFF)
+                            //check for utf-8 surrogate pair
+                            if (character >= 0xD800 && character <= 0xDFFF)
+                            {
+                                for (var k = 0; k < codeLength; ++k)
+                                {
+                                    Skip();
+                                }
+
+                                if (analyzer.Peek(0) == '\\' &&
+                                    (analyzer.Peek(1) == 'u' || analyzer.Peek(1) == 'U'))
+                                {
+                                    Skip(); //escape character
+                                    if (analyzer.Peek(0) == 'u')
+                                    {
+                                        codeLength = 4;
+                                    }
+                                    else
+                                    {
+                                        codeLength = 8;
+                                    }
+                                    Skip(); //escape code
+
+                                    var lowSurrogate = 0;
+
+                                    // Scan the character value.
+                                    for (var k = 0; k < codeLength; ++k)
+                                    {
+                                        if (!analyzer.IsHex(0))
+                                        {
+                                            throw new SyntaxErrorException(start, cursor.Mark(), "While scanning a quoted scalar, did not find expected hexadecimal number.");
+                                        }
+                                        lowSurrogate = ((lowSurrogate << 4) + analyzer.AsHex(k));
+                                    }
+
+                                    for (var k = 0; k < codeLength; ++k)
+                                    {
+                                        Skip();
+                                    }
+
+                                    character = char.ConvertToUtf32((char)character, (char)lowSurrogate);
+                                }
+                                else
+                                {
+                                    throw new SyntaxErrorException(start, cursor.Mark(), "While scanning a quoted scalar, found invalid Unicode surrogates.");
+                                }
+                            }
+                            else if (character > 0x10FFFF)
                             {
                                 throw new SyntaxErrorException(start, cursor.Mark(), "While scanning a quoted scalar, found invalid Unicode character escape code.");
                             }
+                            else
+                            {
+                                // Advance the pointer.
 
-                            value.Append(char.ConvertFromUtf32(character));
-
-                            // Advance the pointer.
+                                for (var k = 0; k < codeLength; ++k)
+                                {
+                                    Skip();
+                                }
 
-                            for (var k = 0; k < codeLength; ++k)
-                            {
-                                Skip();
                             }
+
+                            value.Append(char.ConvertFromUtf32(character));
                         }
                     }
                     else