fenced_divs with preformatted text (source code)

[lifeisfoo]

I'm using fenced_divs to wrap code blocks in order to match a custom style in the output file (markdown->docx).

::: {custom-style="MyCodeStyle"}
function hello() {  
    console.log("world");  
}
:::

Pandoc reads that block as a paragraph, so spaces are removed.

    {
      "t": "Div",
      "c": [
        ["", [], [["custom-style", "MyCodeStyle"]]],
        [
          {
            "t": "Para",
            "c": [
              { "t": "Str", "c": "function" },
              { "t": "Space" },
              { "t": "Str", "c": "hello()" },
              { "t": "Space" },
              { "t": "Str", "c": "{" },
              { "t": "LineBreak" },
              { "t": "Str", "c": "console.log(" },
              {
                "t": "Quoted",
                "c": [{ "t": "DoubleQuote" }, [{ "t": "Str", "c": "world" }]]
              },
              { "t": "Str", "c": ");" },
              { "t": "LineBreak" },
              { "t": "Str", "c": "}" }
            ]
          }
        ]
      ]
    },

Is there a way to force pandoc fenced_divs to behave like a code block producing an output like this?

    {
      "t": "CodeBlock",
      "c": [
        ["", [], []],
        "function hello() {  \n    console.log(\"world\");  \n}  "
      ]
    }

I saw this solution but I would like to avoid converting all spaces to non breaking ones.