Strict namespace handling

[pgundlach]

The speedata Publisher is currently very namespace ignorant. This is getting better with the new XPath parser (lxpath).

For example: the following data

<bar:data  xmlns:bar="foons">
    <bar:child>
        <bar:sub>sub</bar:sub>
    </bar:child>
</bar:data>

should be parsable in default mode with

<Record element="data">

and in "strict" mode with

<Record element="foo:data">

where the foo-namespace is foons.

This could be a complete layout:

<Layout xmlns="urn:speedata.de:2009/publisher/en"
    xmlns:sd="urn:speedata:2009/publisher/functions/en"
    xmlns:foo="foons">

    <Options namespaces="strict" />

    <Record element="foo:data">
        <PlaceObject>
            <Textblock>
                <Paragraph>
                    <Value>Hello world</Value>
                </Paragraph>
            </Textblock>
        </PlaceObject>

        <ProcessNode select="*:child" />
    </Record>

    <Record element="foo:child">
        <Message select="name()"></Message>
        <PlaceObject>
            <Textblock>
                <Paragraph>
                    <Value select="namespace-uri()" />
                </Paragraph>
            </Textblock>
        </PlaceObject>
    </Record>
</Layout>

should print / output bar:child and and foons