Question regarding StringContainerInterface with XML output

[tao]

I have been using an extension commonmark-hint-extension and run into a little bug or question regarding StringContainerInterface.

The class Hint extends StringContainerInterface like this:

// Hint.php
<?php

namespace Ueberdosis\CommonMark;

use League\CommonMark\Node\Block\AbstractBlock;
use League\CommonMark\Node\StringContainerInterface;

class Hint extends AbstractBlock implements StringContainerInterface
{

I am using the example from the plugin README file:

:::important Warning!
This is how the **Markdown** looks.
:::

However when I render this as XML it outputs the text twice:

"""
  <?xml version="1.0" encoding="UTF-8"?>
  <document xmlns="http://commonmark.org/xml/1.0">
      <hint_block heading="Warning!" type="important">This is how the **Markdown** looks.
  
          <text>This is how the </text>
          <strong>
              <text>Markdown</text>
          </strong>
          <text> looks.</text>
      </hint_block>
  </document>
  """

You can see the text This is how the **Markdown** looks exists in plain text and then duplicated in XML text tags. When I remove this interface then it works fine:

"""
  <?xml version="1.0" encoding="UTF-8"?>
  <document xmlns="http://commonmark.org/xml/1.0">
      <hint_block heading="Warning!" type="important">
          <text>This is how the </text>
          <strong>
              <text>Markdown</text>
          </strong>
          <text> looks.</text>
      </hint_block>
  </document>
  """

I don't really understand the purpose of this interface and if it is required and working as expected or if this is a bug in Commonmark or if this is a bug in the package for implementing StringContainerInterface...

Any advice?

[colinodell]

StringContainerInterface should only be applied to leaf nodes - things like Text or HtmlBlock or FencedCode nodes that ONLY contain literal text instead of child nodes.

It looks like this extension is not following that behavior. In fact, HintParser is both storing literal text AND parsing text into child inline nodes. It should not be doing both of these things. This is why you see the text to appear in the rendered XML as both the literal text and those child nodes.