In order to create an element with an attribute and a child, this is
necessary:

    e = Element("foo")
    e.setAttribute("bar", "baz")
    e.appendChild(quux)

It would be preferable if Element.__init__ accepted two additional
parameters to shorten this:

    e = Element("foo", attributes={"bar": "baz"}, children=[quux])

It may also be preferable to have a third new parameter, attributesNS,
to parallel the Element.setAttributeNS method.  This would accept a dict
mapping namespaceURI and qualifiedName to an attribute value.

As you Jean-Paul Calderone seem to know what you're talking about could you provide a patch to get this issue going?  If not I might even have a go myself, even if I do get shot down in flames.

I think this could be sneaked into 3.2 if needed, but is it more work than the benefits actually deliver in the real world?

stupid question but why shouldn't this be possible ? 

class Element(_Element):
	def __init__(self,name,**kwargs):
		_Element.__init__(self,name)
                attributes = kwargs.get("attributes")
                children = kwargs.get("children")
		for key,value in attributes.iteritems():
			self.setAttribute(key,value)
		for child in children:
			self.appendChild(child)

I would tend to disagree with the use of **kwargs. It means the argument list would be implicit, not documented by the code and not checked at runtime. We could end up with anything in there. I'll try to propose a patch tonight that implements Jean-Paul's solution.

Here is my patch for it.

On Jun 2, 2012 7:00 AM, "Alexandre Zani" <report@bugs.python.org> wrote:

>
> Alexandre Zani <alexandre.zani@gmail.com> added the comment:
>
> Here is my patch for it.
>
> ----------
> keywords: +patch
> Added file: http://bugs.python.org/file25796/easier_element_init.patch
>
> _______________________________________
> Python tracker <report@bugs.python.org>
> <http://bugs.python.org/issue4849>
> _______________________________________
>