Can't use XML catalog with XSD files that have <xs:include />

[rnathuji]

While using the XML catalog with XSD configuration with the vscode-xml plugin, we encountered an issue where XSD files that utilize <xs:include /> tags don't work. Take, for example, the following test.xml with test.xsd which includes test-include.xsd:

test.xml

<?xml version="1.0" encoding="UTF-8"?>
<document xmlns="http://foobar.com/test">
  <page>
    <title>Page 1</title>
    <content>Content for page 1</content>
  </page>
  <page>
    <title>Page 2</title>
    <content>Content for page 2</content>
  </page>
</document>

test.xsd

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified" targetNamespace="http://foobar.com/test" xmlns:test="http://foobar.com/test">
  <xs:include schemaLocation="test-include.xsd"/>
</xs:schema>

test-include.xsd

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified" targetNamespace="http://foobar.com/test" xmlns:test="http://foobar.com/test">
  <xs:element name="document">
    <xs:complexType>
      <xs:sequence>
        <xs:element maxOccurs="unbounded" ref="test:page"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
  <xs:element name="page" type="test:page-content"/>
  <xs:complexType name="page-content">
    <xs:sequence>
      <xs:element ref="test:title"/>
      <xs:element ref="test:content"/>
    </xs:sequence>
  </xs:complexType>
  <xs:element name="title" type="xs:string"/>
  <xs:element name="content" type="xs:string"/>
</xs:schema>

With a catalog definition as follows:

test_catalog.xml

<catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog">
  <uri
      name="http://foobar.com/test"
      uri="./test.xsd" />
</catalog>

Opening test.xml results in:

cvc-elt.1.a: Cannot find the declaration of element 'document'.

A "merged" test-merged.xsd, however, works just fine:

test-merged.xsd

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema" elementFormDefault="qualified" targetNamespace="http://foobar.com/test" xmlns:test="http://foobar.com/test">
  <xs:element name="document">
    <xs:complexType>
      <xs:sequence>
        <xs:element maxOccurs="unbounded" ref="test:page"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
  <xs:element name="page" type="test:page-content"/>
  <xs:complexType name="page-content">
    <xs:sequence>
      <xs:element ref="test:title"/>
      <xs:element ref="test:content"/>
    </xs:sequence>
  </xs:complexType>
  <xs:element name="title" type="xs:string"/>
  <xs:element name="content" type="xs:string"/>
</xs:schema>

It seems this may be due to the way resolution occurs in XMLCatalogResolver based upon the namespace when invoked here. As a quick test to help confirm the root cause, we did find that making a similar change as the one in this related SO post to resolveEntity in XMLCatalogResolverExtension.java seemed to correct the problem.

[angelozerr]

@rnathuji thank a lot for reporting this issue and give us a very detailed information.

Is there any chance to have a PR and tests with your suggestion fix?

[rnathuji]

@angelozerr - Sure, I'd be happy to help with a PR. I may, however, need some guidance due to my lack of familiarity 😅 . I've created a draft PR for the time being which contains the quick and dirty change that allowed things to work while debugging / testing. I'd definitely welcome feedback as far as:

1. Whether this is even the proper place to try and hook this logic in
2. The use of resourceIdentifier.getExpandedSystemId() != null as a conditional gate -- It feels brittle to me in that I don't know what underlying assumptions that makes

I can also start looking at the test additions, but LMK if there are any major course corrections that should be made in terms of the implementation approach.