Handling links for headers

[Silverblix]

The following code....

from markdownify import markdownify, ATX

clean_html = '''
<a href="https://www.google.com">
<h4>Google Search</h4>
</a>
'''

md_content = markdownify(clean_html, heading_style=ATX)

print(md_content)

generates...

  [#### Google Search](https://www.google.com) 

It should generate the following MD instead, isn't it?

#### [Google Search](https://www.google.com)

Also, there are 2 white spaces in front that are unnecessary which confuses reliable MD readers.

Fixing the unnecessary white spaces in front of headers in ATX style and a better handling of the <a> tag would allow to generate much higher quality MD output for foundational html tag structures.

Thanks!

Sharing a good reference for me: Markdown Reference Guide

[AlexVonB]

Hi!

The whitespace issue is handled in #17 and soon to be closed, the other issue is interesting. Headings are block elements, links are inline. It is best practice not to embed block elements into inline elements. Anchors should not include headings. I can have a look into how simple it would be to make the other way around happen, but I cannot promise a good solution.

Best! Alex

[Silverblix]

Hi! Thank you for the quick reply.

Which release should include the fix #17 for the whitespace issue?

My observation is that web browsers do render headers (h1, h2, etc.) with links (<a><h1>My Header</h1></a>) properly. I am finding modern websites using headers with links that are not anchors (#...).

[AlexVonB]

Hi, the fix of #17 was released in 0.6.3.

Regarding the a/h1 issue: for this we would have to do what browsers do, they read the HTML and apply a multitude of error correction, so that non-standard HTML gets rendered to the best of their abilities. The parser used in this project tries to do this, too, but it seems that this exact issue is not addressed. So we would have to change the DOM tree for all inlines that contain a block -- or we hardcode the exception for a/h1.