Make HTML reader more forgiving of bad HTML.

* Skip spaces after <b>, <emph>, etc.
* Convert Plain elements into Para when they're in a list
  item with Para, Pre, BlockQuote, CodeBlock.

An example of HTML that pandoc handles better now:

~~~~
<h4> Testing html to markdown </h4>
<ul>
<li>
<b> An item in a list </b>
<p> An introductory sentence.
<pre>
Some preformatted text
at this stage comes next.

But alas! much havoc
is wrought by Pandoc.
</pre>
</ul>
~~~~

Thanks to Dirk Laurie for reporting the issues.

0 files changed, 0 insertions, 0 deletions