Retain whitespace when parsing pure links

[sionide21]

Fixes #22

[sionide21]

While working on this I discovered an issue that may be hiding problems in other tests. It appears that Floki drops whitespace only text nodes.

This means that the spec following spec would have actually passed before this change.

markdown = "**Test**     https://www.google.com"
html = "<p><strong>Test</strong>     <a href=\"https://www.google.com\">https://www.google.com</a></p>\n"
ast      = parse_html(html)
messages = []

assert as_ast(markdown) == {:ok, ast, messages}

[RobertDober]

Thank you for all of this, not easy to understand my code, unfortunately.

I'll check the Floki issue and rewrite the test accordingly, I shall not accept a false negative!

Oh you fixed that 👏

[RobertDober]

I have checked it out into a feature branch for the issue at hand, be not alarmed not to find your code in master yet