Links wrapped in HTML tags generate broken markup with the new link parser

[gocom]

There is some inconsistencies and possible regression with link parsing. Before the new link parser this:

<span>"link":http://example.com</span>

Generated:

<p><span>“link”:http://example.com</span></p>

Now, it generates a link, but also broken HTML:

<p><span><a href="http://example.com&lt;/span%3E">link</a></p>

The link eats the closing tag. The new parser doesn't detect link end if there is no punctuation mark, end of document or space at the end. It holds true for instance when mixing Textile spans and HTML around the link. This:

??"link":http://example.com??

<span>??example??</span>

Generates:

<p><cite><a href="http://example.com">link</a></cite></p>

<p><span><cite>example</cite></span></p>

While this:

<span>??"link":http://example.com??</span>

Generates:

<p><span>??<a href="http://example.com??&lt;/span&gt;">link</a></p>

Generating both broken HTML, improperly encoded URL and leaving raw Textile to the document.

[gocom]

Closed by d79ef7c