XmlMarkupFormatter and self-closing tags

[duncanawoods]

I can't see a way to choose self-closing tags.

The default format of many partially-human readable xml files e.g. csproj files, is to use self-closing tags. If you round trip a file through XmlMarkupFormatter then they will become uglified.

I would appreciate an option to choose self-closing but ideally I would support changing the default to be self-closing because it produces shorter and more readable output and I have no use-case for the verbose form.

Here is my current work-around to wrap the XmlMarkupFormatter:

public class SelfClosingXmlMarkupFormatter : IMarkupFormatter
{
    public string Text(ICharacterData text) =>
        XmlMarkupFormatter.Instance.Text(text);

    public string Comment(IComment comment) =>
        XmlMarkupFormatter.Instance.Comment(comment);

    public string Processing(IProcessingInstruction processing) =>
        XmlMarkupFormatter.Instance.Processing(processing);

    public string Doctype(IDocumentType doctype) =>
        XmlMarkupFormatter.Instance.Doctype(doctype);

    public string OpenTag(IElement element, bool selfClosing) =>
        XmlMarkupFormatter.Instance.OpenTag(element, !(element.HasChildNodes || element.HasTextNodes()));

    public string CloseTag(IElement element, bool selfClosing) =>
        XmlMarkupFormatter.Instance.CloseTag(element, !(element.HasChildNodes || element.HasTextNodes()));

    public string Attribute(IAttr attribute) =>
        XmlMarkupFormatter.Instance.Attribute(attribute);
}

[FlorianRappl]

Sounds good.

Can you share an example how you produced the original format - which is not round-trip preserved? Thanks!

[duncanawoods]

Here you go. Ideally it wouldn't be ToHtml but I understand the naming difficulty across different markup formats. You nicely solve it in the library name AngleSharp... so maybe ToAngles?! Ok maybe not... ToText, ToFormat, ToMarkup would work.

[Test]
public void test_xml()
{
    var xmlDoc = new XmlParser().ParseDocument(
        @"<Project Sdk=""Microsoft.NET.Sdk"">
            <ItemGroup>
                <PackageReference Include=""AngleSharp"" Version=""0.12.1"" />
                <PackageReference Include=""AngleSharp.Xml"" Version=""0.12.1"" />
                <PackageReference Include=""AngleSharp.XPath"" Version=""1.1.4"" />
            </ItemGroup>   
        </Project>");

    {
        var sw = new StringWriter();
        xmlDoc.ToHtml(sw, new XmlMarkupFormatter());
        
        Console.WriteLine("BOO");
        Console.WriteLine(sw.ToString());
        Console.WriteLine();
    }
    
    {
        var sw = new StringWriter();
        xmlDoc.ToHtml(sw, new SelfClosingXmlMarkupFormatter());
        
        Console.WriteLine("YAY");
        Console.WriteLine(sw.ToString());
        Console.WriteLine();
    }
}

[FlorianRappl]

I mean the most ideal would be ToString (and I think it was our original design choice), but it comes with some drawbacks (most notably the debug experience - despite some attributes that can then help out / alleviate the situation a bit).

Independent of the name - thanks for the example; I'll look into it!

[duncanawoods]

Yep, I've experienced the same drawbacks. They should call it ToDebugSummary given it's now so intimately tied into to debuggers/IDEs. My default is ToText for non-trivial string generation which I would otherwise call ToString.

[FlorianRappl]

Part of it landed in devel. Notably:

• You can know configure the XmlMarkupFormatter with an option called IsAlwaysSelfClosing. Setting this property to true will serialize empty elements as self-closing.

What we miss so far is that an element that was declared as self-closing should be serialized as self-closing. For this, however, we require AngleSharp 0.14 - with a breaking change regarding the element construction from documents (allowing to set the respective NodeFlags).

[FlorianRappl]

Fully landed in devel using AngleSharp v0.14 preview.