JsonDiffPatch().Diff()对比含邮箱长字符串异常问题排查

我用JsonDiffPatch().Diff()方法对比两个JSON字符串：

var oldJson = {"Author":"admin","Description":"","LastEditedBy":"admin","LastModifiedDate":"8/3/2023 8:53:09 AM","Name":"Export-5","RecipientEmails":"aa@1.com,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co"};

var currentJson = {"Author":"admin","Description":"","LastEditedBy":"admin","LastModifiedDate":"8/3/2023 8:58:14 AM","Name":"Export-5","RecipientEmails":"aa@1.co,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co"}

JToken diff = new JsonDiffPatch().Diff(JToken.Parse(oldJson), JToken.Parse(currentJson));

不含邮箱列表时对比结果正常，但包含邮箱列表时，得到的差异如下：

"LastModifiedDate": [
  "8/3/2023 8:53:09 AM",
  "8/3/2023 8:58:14 AM"
]
"RecipientEmails": [
  "@@ -1,20 +1,19 @@\n aa@1.co\n-m\n ,a@1.com,aaq\n",
  0,
  2
]

为什么会出现多余字符？怎么去除这些字符并得到标准的差异数组？完整代码示例：

using System;
using JsonDiffPatchDotNet;
using Newtonsoft.Json;
using Newtonsoft.Json.Linq;

public class Program
{
    public static void Main()
    {
        Console.WriteLine("Hello World");
        string oldJson = "{\"RecipientEmails\":\"aa@1.com,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co\"}";
        string currentJson = "{\"RecipientEmails\":\"aa@1.co,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co\"}";
        JToken diff = new JsonDiffPatch().Diff(JToken.Parse(oldJson), JToken.Parse(currentJson));
        Console.WriteLine(diff);
    }
}

JsonDiffPatch默认对字符串采用文本差异算法（类似Git的diff格式），当字符串内容较长或被识别为可拆分的文本时，会生成这种带@@标记的diff格式，而非直接返回新旧字符串的数组。这是库默认启用的文本差异优化，目的是针对长字符串只输出变化部分，减少差异结果的体积。

要让字符串对比返回标准的新旧值数组，只需修改JsonDiffPatch的配置，关闭文本差异功能，强制对所有字符串使用简单的全量对比。

修改后的代码如下：

using System;
using JsonDiffPatchDotNet;
using Newtonsoft.Json;
using Newtonsoft.Json.Linq;

public class Program
{
    public static void Main()
    {
        Console.WriteLine("Hello World");
        string oldJson = "{\"RecipientEmails\":\"aa@1.com,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co\"}";
        string currentJson = "{\"RecipientEmails\":\"aa@1.co,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co\"}";
        
        // 配置JsonDiffPatch，禁用文本差异算法
        var jdp = new JsonDiffPatch(new JsonDiffPatchOptions
        {
            TextDiff = TextDiffMode.Disabled
        });
        
        JToken diff = jdp.Diff(JToken.Parse(oldJson), JToken.Parse(currentJson));
        Console.WriteLine(diff);
    }
}

修改后，RecipientEmails的差异会变成标准的数组格式：

"RecipientEmails": [
  "aa@1.com,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co",
  "aa@1.co,a@1.com,aaq@1.com,ad@1.com,af@1.com,a@1a.com,a@1v.com,za@1.com,ra@1.com,a@11.co"
]

---
内容的提问来源于stack exchange，提问作者tangel