Let XPathContext correctly pick the default namespaceURI

[GoogleCodeExporter]

[Description]

Since XPathContext tries to retrieve the default namespace URI from the 
document elements' xmlns attribute, it doesn't work in the following situation 
where a) xmlns attribute is modified afterward b) the default namespace URI for 
the document without namespace specifications is implied by the DOM parser.

I was stuck on this in the case (b) because I have a DOM implementation other 
than minidom (actually backed by the WebKit DOM through PyObjC).

Patch attached.

[Code for reproduction]

from xml.dom.minidom import parseString
from xpath import XPathContext

doc = parseString('''
<foo xmlns="http://foo.example.com/">
  <item>foo</item>
  <item xmlns="http://bar.example.com/">bar</item>
</foo>
''')
doc.documentElement.setAttribute('xmlns', 'http://bar.example.com/')
assert doc.documentElement.namespaceURI == 'http://foo.example.com/'
assert doc.documentElement.getAttribute('xmlns') == 'http://bar.example.com/'
nodeset = XPathContext(doc).find('//item', doc)
assert len(nodeset) == 1
assert nodeset[0].namespaceURI == 'http://foo.example.com/' # boom!


[What version of the product are you using? On what operating system?]

SVN head

Original issue reported on code.google.com by m...@mozo.jp on 6 Jan 2013 at 12:09

Attachments:

• py-dom-xpath-xpathctxt-namespaceuri.20130106.patch.diff