Tag object lost some by html5lib with incorrect html

[Hara Hidekazu]

I faced one problem which Tag object lost some information by html5lib with incorrect html & after losing them I'll fail to operate it as a BS4 object.

When I try to operate following soup_html5lib obj from html1, I can do soup_html5lib.div , but I failed soup_html5lib.div.find_all('img').
I tried to find difference between a soup_lxml and soup_html5lib.

I found a point which 'parser_class' is lost in Tag obj with html5lib.

I know html1 has wrong html structure. But I can make Tag object once.

I'm not sure this is BS4 bug or html5lib or other.

I just expect avoid Traceback before operate.
If there is some option for avoiding it , let's me know.

Thanks for reading.

-----
from bs4 import BeautifulSoup
html1 = '''
<div><a href="http://www.theborneo.com/B16092503.jpg"><img src="http://www.theborneo.com/B16092503.jpg" /><p><noscript><img src="http://www.theborneo.com/B16092503.jpg" /></noscript></a> The woman mourning</p></div>
'''
html2 = '''
<div><a href="http://www.theborneo.com/B16092503.jpg"><img src="http://www.theborneo.com/B16092503.jpg" /><p><noscript><img src="http://www.theborneo.com/B16092503.jpg" /></noscript></p></a> The woman mourning</div>
'''

soup_lxml = BeautifulSoup(html1, 'lxml')
soup_lxml.div.find_all('img')

soup_html5lib = BeautifulSoup(html1, 'html5lib')
soup_html5lib.div.find_all('img')
...Traceback...
AttributeError: 'NoneType' object has no attribute 'next_element'
...

vars(soup_html5lib.div)
...
 'parser_class': None,
...
-----

on mysystem:
- Ubuntu 14.04.5 LTS
- Python 2.7.6
on pip:
- ipython (1.2.1)
- beautifulsoup4 (4.5.1)
- lxml (3.3.3)
- html5lib (0.999999999)

[Leonard Richardson (leonardr)]

Fixed in revision 434.