Fix xsd:extension to extend the base type with its elements.

[ph777]

Consider the following schema that extends complex types:

<xs:complexType name="baseType">
        <xs:sequence>
            <xs:element name="a" type="xs:string"/>
        </xs:sequence>
    </xs:complexType>

    <xs:complexType name="extendedType">
        <xs:complexContent>
            <xs:extension base="baseType">
                <xs:sequence>
                    <xs:element name="b" type="xs:string"/>
                </xs:sequence>
            </xs:extension>
        </xs:complexContent>
    </xs:complexType>

    <xs:complexType name="superExtendedType">
        <xs:complexContent>
            <xs:extension base="extendedType">
                <xs:sequence>
                    <xs:element name="c" type="xs:string"/>
                </xs:sequence>
                <xs:attribute name="attr" type="xs:string"/>
            </xs:extension>
        </xs:complexContent>
    </xs:complexType>

    <xs:element name="myElem" type="superExtendedType"/>

According to schema in elemenet myElem elements a, b, c can be used. This PR fixes autocompletion to correctly offer all of them.

Please review the PR and consider pulling it to your master repo.

[ph777]

Hello, do you plan to include this (and the following PRs #47 and #48) into your project?

If you think there is an error in the implementation or another problem, please let me know and I'll be happy to fix it. So far I think all the PRs improve your package in terms of better support for XSD constructs and I'll be pleased if you merge them in.

[pleonex]

Yes, thank you so much for the contributions!!
I just want to review the PR and test it before publishing the new version. Hopefully this weekend I will find some time. I'll let you know with the feedback (if any).

[ph777]

That is great news. Thanks for it!

Let me know if help or changes needed.

[pleonex]

Thanks for the changes!
My only concern here is that we are assuming here that when we are post-processing superExtendedType, we have already post-processed extendedType so it contains the elements from baseType, right?

If I am not wrong, this may not be the case always since we are iterating a dictionary, so the order of post-processing is a bit random. But your changes are good for adding to the extension the node elements. Maybe we can address that in a different issue.

[ph777]

Yes, exactly. This simple implementation of xs:extension works only for XSDs where the definition of the baseType precedes the definition of the extended type. However, it is still an improvement for well-ordered XSDs.

To handle also XSDs with xs:extension definition preceding its base type, there would have to be more then a single loop through types in post-processing. A kind of a stack with unresolved extensions would be required to keep track of types that still wait for resolution of their base type. I aggree that this can be handled as a different issue.

[pleonex]

I have some ideas for the multi round post-processing, it will be implemented from #50. Thanks!