Un-escape specific characters when serializing to JSON in C#

Question

I have a simple task which is to unescape an apostrophe when serializing to JSON using System.Text.JSON:

using System;
using System.Text.Encodings.Web;
using System.Text.Json;
using System.Text.Unicode;

namespace DebugApp;

internal class Program
{
    private static void Main(string[] args)
    {
        var input = new Item
        {
            Text = "'"
        };

        var encoderSettings = new TextEncoderSettings();
        encoderSettings.AllowCharacters('\u0027', '\'');
        encoderSettings.AllowRange(UnicodeRanges.BasicLatin);
        var options = new JsonSerializerOptions
        {
            Encoder = JavaScriptEncoder.Create(encoderSettings)
        };

        var json = JsonSerializer.Serialize(input, options);

        Console.WriteLine("Serialized json: {0}", json);
    }

    private class Item
    {
        public string Text { get; set; }
    }
}

I expect the outcome to be Serialized json: {"Text":"'"}

Unfortunately, I am still getting Serialized json: {"Text":"\u0027"}

I referred to the Microsoft documentation here => Serialize specific characters

What am I doing wrong?

Answer 1

Use Newtonsoft.Json and you always will be happy. But if you still want to use System.Text.Json , use this System.Text.RegularExpressions function

json = System.Text.RegularExpressions.Regex.Unescape(
         System.Text.Json.JsonSerializer.Serialize(input));

or more simple

 json = System.Text.Json.JsonSerializer.Serialize(input).Replace("\\u0027","'");

result

{"Text":"'"}

Answer 2

Unfortunately this is not supported in the current version of .NET.

You could do this with your own encoder, but that would require rewriting huge chunks of code.

---

Instead, you can do this with some rather nasty reflection. Please note that this code is quite brittle and likely to break on a different version of .NET. It doesn't work on .NET 4, as the internal code is different.

 private static void Main(string[] args)
 {
     var input = new Item
     {
         Text = "'"
     };

     var encoderSettings = new TextEncoderSettings();
     encoderSettings.AllowRange(UnicodeRanges.BasicLatin);
     var encoder = System.Text.Encodings.Web.JavaScriptEncoder.Create(encoderSettings);
     
     var innerEncoder = encoder.GetType().GetField("_innerEncoder", BindingFlags.NonPublic | BindingFlags.Instance).GetValue(encoder);
     var codePoints = innerEncoder.GetType().GetField("_allowedBmpCodePoints", BindingFlags.NonPublic | BindingFlags.Instance).GetValue(innerEncoder);
     codePoints.GetType().GetMethod("AllowChar").Invoke(codePoints, new object[]{ '\''});
     // codePoints is a struct so need to write it back
     innerEncoder.GetType().GetField("_allowedBmpCodePoints", BindingFlags.NonPublic | BindingFlags.Instance).SetValue(innerEncoder, codePoints);

     var options = new JsonSerializerOptions
     {
         Encoder = encoder,
     };

     var json = JsonSerializer.Serialize(input, options);

     Console.WriteLine("Serialized json: {0}", json);
 }

 private class Item
 {
     public string Text { get; set; }
 }

dotnetfiddle