Fix XML namespaces wrongly using https

[gnodet]

Description:
This PR fix the XML namespaces used when creating the maven toochains.xml file.
The namespaces are identifiers and do not start with 'https', but 'http'.
See https://www.w3.org/TR/xmlschema-1/#Instance_Document_Constructions and
https://maven.apache.org/xsd/toolchains-1.1.0.xsd.
This is problematic and will break when Maven 4 uses a more conformant XML parser.

Related issue:
Add link to the related issue.

Check list:

• Mark if documentation changes are required.
• Mark if tests were added or updated to cover the changes.

[gnodet]

The outcome is the following failure: https://github.com/apache/maven/actions/runs/5398872373/jobs/9806118625?pr=1185#step:6:12518

Warning: 
Warning:  Some problems were encountered while building the effective toolchains
Warning:  ParseError at [row,col]:[2,1]
Message: Unknown attribute 'schemaLocation' for tag 'toolchains' @ line 2, column 1
Warning: 

[IvanZosimov]

Hi, @gnodet 👋 Are you sure that https scheme can't be used and the root cause of your problem is there? If you take a look at the https://www.w3.org/TR/xmlschema-1/#xsi.schemaLocation you may see that it can be of type https://www.w3.org/TR/2004/REC-xmlschema-2-20041028/datatypes.html#anyURI. As the anyURI represents a Uniform Resource Identifier Reference, we can find information about URI reference in https://www.ietf.org/rfc/rfc2396.txt. It seems like this standard doesn't limit the number of URI schemas to those listed in the standard, like the http. The https is just yet another standardised https://datatracker.ietf.org/doc/html/rfc7230#section-2.7.2 URI schema. Correct me if I'm wrong, please.

[IvanZosimov]

Just a gentle ping, @gnodet 📟

[gnodet]

Hi, @gnodet 👋 Are you sure that https scheme can't be used and the root cause of your problem is there? If you take a look at the https://www.w3.org/TR/xmlschema-1/#xsi.schemaLocation you may see that it can be of type https://www.w3.org/TR/2004/REC-xmlschema-2-20041028/datatypes.html#anyURI. As the anyURI represents a Uniform Resource Identifier Reference, we can find information about URI reference in https://www.ietf.org/rfc/rfc2396.txt. It seems like this standard doesn't limit the number of URI schemas to those listed in the standard, like the http. The https is just yet another standardised https://datatracker.ietf.org/doc/html/rfc7230#section-2.7.2 URI schema. Correct me if I'm wrong, please.

@IvanZosimov
Any valid URI could be used to define a schema URI, you're absolutely right. However, in order to be a valid maven toolchains xml file, the one defined by the Maven schema MUST be used, else the xml simply refers to a different schema namespace. The same is true for the xsi namespace. So the namespace identifiers must match those defined by the schemas and if you look at https://www.w3.org/TR/xmlschema-1/#Instance_Document_Constructions and
https://maven.apache.org/xsd/toolchains-1.1.0.xsd, the namespaces are identified with the http protocol in their URI. That's unrelated to the actual schema location (inside the xsi:schemaLocation) which is a URL that just needs to point to the actual xml schema, and thus could use http or https.

Please try to validate the following file with an online xml validator such as https://www.freeformatter.com/xml-validator-xsd.html, you should end up with an error Cvc-elt.1.a: Cannot Find The Declaration Of Element 'toolchains'., Line '3', Column '116'. because the toolchains element is not part of the https://maven.apache.org/TOOLCHAINS/1.1.0 namespace, but the http://maven.apache.org/TOOLCHAINS/1.1.0

<?xml version="1.0"?>
<toolchains xmlns="https://maven.apache.org/TOOLCHAINS/1.1.0"
  xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="https://maven.apache.org/TOOLCHAINS/1.1.0 https://maven.apache.org/xsd/toolchains-1.1.0.xsd">
  <toolchain>
    <type>jdk</type>
    <provides>
      <version>17</version>
      <vendor>Eclipse Temurin</vendor>
      <id>temurin_17</id>
    </provides>
    <configuration>
      <jdkHome>/opt/hostedtoolcache/Java_Temurin-Hotspot_jdk/17.0.1-12/x64</jdkHome>
    </configuration>
  </toolchain>
</toolchains>

Using the fixed xml below pass the validation successfully:

<?xml version="1.0"?>
<toolchains xmlns="http://maven.apache.org/TOOLCHAINS/1.1.0"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="http://maven.apache.org/TOOLCHAINS/1.1.0 https://maven.apache.org/xsd/toolchains-1.1.0.xsd">
  <toolchain>
    <type>jdk</type>
    <provides>
      <version>17</version>
      <vendor>Eclipse Temurin</vendor>
      <id>temurin_17</id>
    </provides>
    <configuration>
      <jdkHome>/opt/hostedtoolcache/Java_Temurin-Hotspot_jdk/17.0.1-12/x64</jdkHome>
    </configuration>
  </toolchain>
</toolchains>

[gnodet]

Fwiw, maven 3 does not validate the XML when they are loaded. That's why this issue has not caused any real issue at this point. However maven 4 will reject any unknown attribute, hence this PR to fix the problem before it actually hits people.

[gnodet]

To be more precise, the error I mentioned above from the validator is related to the fact that the xsi:schemaLocation is simply ignored because xsi points to the wrong namespace.

If we fix the xsi namespace with the below file, a bunch of different and more precise errors will be printed: TargetNamespace.1: Expecting Namespace 'https://maven.apache.org/TOOLCHAINS/1.1.0', But The Target Namespace Of The Schema Document Is 'http://maven.apache.org/TOOLCHAINS/1.1.0'., Line '24', Column '195'. because of the mismatch between the defined namespace and the one that is actually used.

<?xml version="1.0"?>
<toolchains xmlns="https://maven.apache.org/TOOLCHAINS/1.1.0"
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  xsi:schemaLocation="https://maven.apache.org/TOOLCHAINS/1.1.0 https://maven.apache.org/xsd/toolchains-1.1.0.xsd">
  <toolchain>
    <type>jdk</type>
    <provides>
      <version>17</version>
      <vendor>Eclipse Temurin</vendor>
      <id>temurin_17</id>
    </provides>
    <configuration>
      <jdkHome>/opt/hostedtoolcache/Java_Temurin-Hotspot_jdk/17.0.1-12/x64</jdkHome>
    </configuration>
  </toolchain>
</toolchains>

[IvanZosimov]

Huge thanks for the clarification, @gnodet ❤️