Re: [xslt] extension function, python, nodeset ?

[Frédéric Glorieux <fred fictif org>]

Hi,

Sorry to answer to myself, I got something which seems to work, but it 
sure need improvement from better python/libxml gurus.

def test(ctx):
  """
  XSLT extension function, a skeleton for regexp tagging of a string
  """
  try:

    # string tagging ...

    # is it encoding safe ?
    result="1 <é>2</é> 3"

    doc = libxml2.parseDoc("<root>" + result + "</root>")
    root= doc.getRootElement()

    # sometimes needed, but can't say when
    # root.unlinkNode()
    # OK, but with root element
    # return [root]

    # to return a node list, is there better in  API ?
    nl = []
    child = root.children
    while child is not None:
      nl.append(child)
      child = child.next
      # unlink needed, or you can get strange results
      # like mixing, "1 <é>2</é> 3" => "<é>2</é> 31 "
      # but do it carefully, or child will loose its next
      nl[len(nl) - 1].unlinkNode()

    # seems OK
    return nl
  finally:
    doc.freeDoc()


--
Frédéric Glorieux <http://fictif.org>