Missing and null values

The HAR format defines which fields are required, but the tools that produce archives do not always follow it: an archive can hold an explicit null for any field. So that those archives still parse, every reference property of the model is nullable, and it is up to you to decide what an absent value means.

A field that is missing from the JSON and a field that is null are read differently:

var document = HarDocument.Parse("""
    {
      "log": {
        "entries": [
          { "request": { "method": "GET", "url": "https://example.com/", "headers": null } },
          { "request": null }
        ]
      }
    }
    """);

var entries = document.Log!.Entries!;

// A missing property keeps its default value
Assert.Equal("", entries[0].Request!.HttpVersion);
Assert.Empty(entries[0].Request!.Cookies!);

// An explicit null is kept as null
Assert.Null(entries[0].Request!.Headers);
Assert.Null(entries[1].Request);

• A missing field keeps the default value of the property: "" for most strings, an empty list for the collections, an empty object for HarEntry.Request, Response, Cache and Timings, 1.2 for HarLog.Version.
• An explicit null gives null, even for a collection or an object.
• The optional fields of the format, such as Comment, PostData or HarLog.Pages, are null when they are missing.

When you read an archive you did not create, check for null with ?. and ??, as the examples of this documentation do. When you write an archive, a null property is left out of the JSON.

ToHttpRequestMessage and ToHttpResponseMessage accept these values: null headers, cookies or MIME types give a message without them. They throw an InvalidOperationException when the entry has no request or no response, or when the method or the URL of the request is null. A request also needs a method and an absolute URL: an empty method, or a URL that Uri cannot parse, throws like new HttpMethod and new Uri do. See Converting entries.