Markdown_to_html cannot handle HTML elements

[jannekalliola]

Describe the bug

The markdown_to_html filter does not transform Markdown if it contains HTML elements.

To reproduce

Have Markdown code with HTML in a variable:

# Privacy Policy
 
<article class="directtext">

Some text here

</article>

Try to render it in a template

{{ mdvariable | markdown_to_html }}

Markdown code before the first HTML element is rendered nicely and after that there is only Markdown source code.

Expected behavior

The filter should render the whole Markdown content as MD specification allows inclusion of HTML tags.

Version

8.7.9

Operating System

No response

Relevant log output

[jannekalliola]

Found out after some further testing that Markdown code inside HTML tags is not rendered properly in normal files, either.

[ArnaudLigny]

Hum, may be caused by the last release of Parsedown. I should do more tests...
Thanks for the report I will check that.

[ArnaudLigny]

@jannekalliola I see that you are using an "old" version of Cecil ("8.7.9"), would you be able to update it?

[ArnaudLigny]

After doing some research, I'm wondering if this isn't actually how things work: if the content of the variable filtered with markdown_to_html is indented, then it will be rendered as raw code.

Could you check this?

ref: https://www.markdownguide.org/basic-syntax/#code-blocks

[jannekalliola]

@ArnaudLigny I typoed the version number, it was 8.79. I've now tested it with 8.88 and had the same issue.

This is not related to code blocks and it is not intended in the file. I've attached an example Markdown file.

cookies.md

[jannekalliola]

I downloaded the latest version of Parsedown and ParsedownExtra. Made a simple test script (attached). It seems that Parsedown was able to render MD code inside HTML elements, but ParsedownExtra could not.

pd.php

[jannekalliola]

I checked the differences of those two libraries and it boiled down to how interrupts are handled in method blockMarkupContinue in ParsedownExtra:

When I made the following change, it started to work as expected:

if (isset($Block['interrupted']))
{
    $Block['element']['rawHtml'] .= "\n";
    unset($Block['interrupted']);
    return;
}

I added the empty return statement on row 314, based on how Parsedown blockMarkupContinue method works.

I unfortunately have no clue whether this change breaks something else, but now the contents of HTML tags are not marked as raw HTML and they are processed by the Markdown processor.

[ArnaudLigny]

Could you try without extra spaces in HTML tags:

# Cookies

<article class="directtext">
This site does not use cookies. Therefore, there is no cookie banner and no tracking.
</article>

instead of :

# Cookies

<article class="directtext">

This site does not use cookies. Therefore, there is no cookie banner and no tracking.

</article>

[jannekalliola]

When I remove empty lines, then both libraries fail to render Markdown snippet between HTML elements.

I'll open a ticket to ParsedownExtra project. In a meanwhile, I need to get my site up and running. The ParsedownToc implementation within Cecil is buried inside the project. Is there any possibility to amend it without modifying Cecil's code? Or should I just open the .phar file and make the change until this bug is fixed?

[jannekalliola]

➡️ erusev/parsedown-extra#198

[ArnaudLigny]

The ParsedownToc implementation within Cecil is buried inside the project. Is there any possibility to amend it without modifying Cecil's code? Or should I just open the .phar file and make the change until this bug is fixed?

If you're in a hurry, the simplest approach in my opinion is to pull Cecil locally, install the dependencies with Composer, modify the relevant PHP files, and then run the build via the bin script:

php bin/cecil build -vv path/to/your/site

[jannekalliola]

Yes, thanks. I exported the files from the phar file and made the modifications. Now it works as I've intended.