Rendering the Markdown

The Markdown that the converter writes follows CommonMark, plus four common extensions: pipe tables, task lists, strikethrough (~~) and definition lists. A renderer that only implements CommonMark shows these as text. With Meziantou.Framework.Markdown, enable them in the pipeline:

var markdown = HtmlToMarkdown.Convert("<table><tr><th>Name</th><th>Price</th></tr><tr><td>Apple</td><td><del>1.50</del> 1.20</td></tr></table>");
Assert.Equal("""
    | Name | Price |
    | --- | --- |
    | Apple | ~~1.50~~ 1.20 |
    """, markdown);

// Pipe tables and strikethrough are extensions of CommonMark: enable them to render the Markdown
var pipeline = new MarkdownPipelineBuilder()
    .UseAdvancedExtensions()
    .Build();
var html = MarkdownConverter.ToHtml(markdown, pipeline);
Assert.Contains("<td><del>1.50</del> 1.20</td>", html, StringComparison.Ordinal);

UseAdvancedExtensions enables all four, with other extensions: see Extensions to enable them one by one.

Round trip

Converting HTML to Markdown and rendering it again usually gives an equivalent HTML for the converted elements: the same structure and the same text. It does not give the same markup:

• the attributes that Markdown cannot write (class, id, style...) are lost, and the containers (div, section...) are replaced by their content;
• the whitespace is collapsed, outside of the code blocks;
• a list can change between tight and loose, so its items gain or lose their p elements;
• the renderer writes its own markup, such as the thead and tbody of a table, or the attributes of a task list checkbox.

The unknown elements written as raw HTML are rendered again as they are, when the renderer allows raw HTML. Read Security before you render Markdown that comes from untrusted HTML.