The htmldiff function fails with a "KeyError" exception when the html contains img tags without a src attribute. Although this is not valid html, I have found that it is a common issue in many sites, so I think that it should be better to ignore these kind of tags.
Here's a simple example:
>>> from lxml.html.diff import htmldiff
>>> htmldiff('<img />', '<img />')
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/lucas/pythonwebalert/local/lib/python2.7/site-packages/lxml-2.3.1-py2.7-linux-x86_64.egg/lxml/html/diff.py", line 168, in htmldiff
old_html_tokens = tokenize(old_html)
File "/home/lucas/pythonwebalert/local/lib/python2.7/site-packages/lxml-2.3.1-py2.7-linux-x86_64.egg/lxml/html/diff.py", line 534, in tokenize
return fixup_chunks(chunks)
File "/home/lucas/pythonwebalert/local/lib/python2.7/site-packages/lxml-2.3.1-py2.7-linux-x86_64.egg/lxml/html/diff.py", line 576, in fixup_chunks
for chunk in chunks:
File "/home/lucas/pythonwebalert/local/lib/python2.7/site-packages/lxml-2.3.1-py2.7-linux-x86_64.egg/lxml/html/diff.py", line 691, in flatten_el
for item in flatten_el(child, include_hrefs=include_hrefs):
File "/home/lucas/pythonwebalert/local/lib/python2.7/site-packages/lxml-2.3.1-py2.7-linux-x86_64.egg/lxml/html/diff.py", line 682, in flatten_el
yield ('img', el.attrib['src'], start_tag(el))
File "lxml.etree.pyx", line 2198, in lxml.etree._Attrib.__getitem__ (src/lxml/lxml.etree.c:49115)
KeyError: 'src'
I'm attaching a patch that solves the issue by ignoring the tag when the src attribute is not found.
Thanks for the report. I committed a fix here:
https:/