[lxml-dev] Bug in lxml.html with nameless form submit?

Mon Feb 11 08:04:11 CET 2008

This is a pretty common idiom in html - to have a submit button that
has no name:

  <input type="submit" />

But lxml.html barfs if you use the form's fields.items() or
fields.values() method:

>>> import lxml.html
>>> tree = lxml.html.fromstring('''
... <html><body>
...  <form>
...   <input name="foo" value="bar"/>
...   <input type="submit" />
...  </form>
... </body></html>
... ''')
>>> tree
<Element html at 2b58b8edea78>
>>> tree.forms
tree.forms
>>> tree.forms[0]
<Element form at 2b58b8edec80>
>>> tree.forms[0].fields
<FieldsDict for form 0>
>>> tree.forms[0].fields.keys()
[None, 'foo']
>>> tree.forms[0].fields.items()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/lib/python2.4/UserDict.py", line 112, in items
    return list(self.iteritems())
  File "/usr/lib/python2.4/UserDict.py", line 101, in iteritems
    yield (k, self[k])
  File "/usr/lib64/python2.4/site-packages/lxml-2.0-py2.4-linux-x86_64.egg/lxml/html/__init__.py", line 749, in __getitem__
    return self.inputs[item].value
  File "/usr/lib64/python2.4/site-packages/lxml-2.0-py2.4-linux-x86_64.egg/lxml/html/__init__.py", line 811, in __getitem__
    raise KeyError(
KeyError: 'No input element with the name None'
>>> tree.forms[0].fields.values()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/lib/python2.4/UserDict.py", line 110, in values
    return [v for _, v in self.iteritems()]
  File "/usr/lib/python2.4/UserDict.py", line 101, in iteritems
    yield (k, self[k])
  File "/usr/lib64/python2.4/site-packages/lxml-2.0-py2.4-linux-x86_64.egg/lxml/html/__init__.py", line 749, in __getitem__
    return self.inputs[item].value
  File "/usr/lib64/python2.4/site-packages/lxml-2.0-py2.4-linux-x86_64.egg/lxml/html/__init__.py", line 811, in __getitem__
    raise KeyError(
KeyError: 'No input element with the name None'

If you give the submit element a name attribute, this doesn't happen.

-- 

Paul Winkler
http://www.slinkp.com