Whitespace in code blocks is not properly preserved

[SteffenDE]

Hi there,

I'm parsing markdown files with earkmark and hooking into the postprocess to highlight the code with makeup. I noticed that the code earmark parsed does not properly preserve whitespace and therefore also the highlighted code doesn't render correctly:

"""
```xml
<!-- Basic element with attributes -->
<book id="b1" category="fiction" inStock="yes">
  <title>The Great Gatsby</title>
  <!-- Multiple elements of same type -->
  <author>F. Scott Fitzgerald</author>
  <published>1925</published>
  <!-- CDATA section for text with special characters -->
  <description><![CDATA[A story about the American Dream & society in the 1920s]]></description>
</book>
```
"""
|> Earmark.as_html!(
  escape: false,
  postprocessor: fn
    {"code", _attrs, [body], _meta} = node ->
      IO.puts(body)
      node    

    node ->
      node
  end
)

This outputs

<!-- Basic element with attributes -->

<book id="b1" category="fiction" inStock="yes">
  <title>The Great Gatsby</title>
<!-- Multiple elements of same type -->

  <author>F. Scott Fitzgerald</author>
  <published>1925</published>
<!-- CDATA section for text with special characters -->

  <description><![CDATA[A story about the American Dream & society in the 1920s]]></description>
</book>

while it should output

<!-- Basic element with attributes -->
<book id="b1" category="fiction" inStock="yes">
  <title>The Great Gatsby</title>
  <!-- Multiple elements of same type -->
  <author>F. Scott Fitzgerald</author>
  <published>1925</published>
  <!-- CDATA section for text with special characters -->
  <description><![CDATA[A story about the American Dream & society in the 1920s]]></description>
</book>

Using earmark 1.4.47.