Backend engine problem, html recognition error of lxml

[caesar0301]

I encountered a problem when extracting html using BS4. I choose backend engine between lxml and html5lib. But they give me different answers. I know its not the problem of BS. Yet it is valuable to record this difference to let others be aware. Here is my code snippet,

#------------------CODE STARTED-----------------------
#!/usr/bin/env python
#coding: utf-8

html_doc = """
<ul class="list-main-icnset">
<li>
<i class="cell maincell">
 <p class="title"><a target="_blank" href="http://www.itjuzi.com/company/42904"><span>hello</span></a></p>
</i>
</li>
</ul>
"""

from bs4 import BeautifulSoup
soup = BeautifulSoup(html_doc, "lxml")

for ul in soup.find_all('ul', class_ = 'list-main-icnset'):
 for li in ul.find_all('li'):
  print li
  for i in li.find_all('i'):
   print("==================")
   print i
#---------------------CODE ENDED----------------------

With lxml, it gives me the <p> tag out of the scope of <li>, actually it should be within the scope:

<li>
<i class="cell maincell">
</i><p class="title"><a href="http://www.itjuzi.com/company/42904" target="_blank"><span>hello</span></a></p>
</li>
==================
<i class="cell maincell">
</i>

With html5lib, it gives me correct answer:

<li>
<i class="cell maincell">
 <p class="title"><a href="http://www.itjuzi.com/company/42904" target="_blank"><span>hello</span></a></p>
</i>
</li>
==================
<i class="cell maincell">
 <p class="title"><a href="http://www.itjuzi.com/company/42904" target="_blank"><span>hello</span></a></p>
</i>

I did not review the implementation of BS4 how it connects with backends. I could just assume it is the flaws of lxml.

[Leonard Richardson (leonardr)]

Thanks for taking the time to post an example of the differences between parsers where people can see it. Since as you mention this isn't a bug in Beautiful Soup, I'm going to close this issue.