Fix gradle build files parsing issues

[thewolt]

What's changed?

Improved the groovy parser to support:

• reference to FQNC like in the ProcessBuilder.Redirect.to(..) call
• groovy xml node attribute manipulation like n.@conf = 'runtime->default;docs->docs;sources->sources'

What's your motivation?

I tried to use the ChangeDependency recipe on an old gradle build with lots of custom code

Anything in particular you'd like reviewers to focus on?

• in GroovyTypeMapping#classType, in production i've encountered some ClassCastException where JavaType.Unknown cannot be casted to JavaType.Class but i was not able to reproduce in unit tests. I found that downcasting to JavaType.FullyQualified avoids the CCE and does not seem to break something else

Anyone you would like to review specifically?

Have you considered any alternatives or workarounds?

Any additional context

Checklist

• I've added unit tests to cover both positive and negative cases
• I've read and applied the recipe conventions and best practices
• I've used the IntelliJ IDEA auto-formatter on affected files

[timtebeek]

Thanks for diving in for a fix! Let us know when you'd consider this ready for review.

[thewolt]

I've got one question though... it seems that the issue i'm trying to fix is not related to the ChangeDependency recipe specifically but rather the gradle parser in general.

Is there a better place to put the test?

[timtebeek]

Indeed looks like you're fixing a parser bug instead; that should help even more folks & cases, thanks! Might be easier to reproduce in a smaller test as seen in these tree test examples. Does that help?

[thewolt]

Yes it absolutely helped. The tests have been successfully moved there and I may have found another bug when an inner class is imported.

The last thing is about the FIXME that i left in GroovyTypeMapping... it is a private method and the signature change does not seem to bother the two usages. Unfortunately I was not able to reproduce it in unit test. Does this change require a unit test ?

[timtebeek]

Glad to hear that helped! The new smaller tests look much cleaner indeed. I'm tagging Knut or Shannon for a review here, as they have more experience with the Groovy/Gradle parsers, and are more likely to spot any possible impact I might overlook. I'd think it's fine not to add the explicit test you mentioned if it's very awkward to trigger or test for, but perhaps they feel differently.

[thewolt]

Build failed because:

Caused by:
                    org.gradle.internal.resolve.ModuleVersionNotFoundException: Could not find org.openrewrite:rewrite-gradle:8.19.0-SNAPSHOT.
                    Searched in the following locations:
                      - file:/home/runner/.m2/repository/org/openrewrite/rewrite-gradle/8.19.0-SNAPSHOT/maven-metadata.xml
                      - file:/home/runner/.m2/repository/org/openrewrite/rewrite-gradle/8.19.0-SNAPSHOT/rewrite-gradle-8.19.0-SNAPSHOT.pom
                      - https://oss.sonatype.org/content/repositories/snapshots/org/openrewrite/rewrite-gradle/8.19.0-SNAPSHOT/maven-metadata.xml
                      - https://oss.sonatype.org/content/repositories/snapshots/org/openrewrite/rewrite-gradle/8.19.0-SNAPSHOT/rewrite-gradle-8.19.0-20240227.124952-11.pom
                      - https://repo.maven.apache.org/maven2/org/openrewrite/rewrite-gradle/8.19.0-SNAPSHOT/maven-metadata.xml
                      - https://repo.maven.apache.org/maven2/org/openrewrite/rewrite-gradle/8.19.0-SNAPSHOT/rewrite-gradle-8.19.0-SNAPSHOT.pom
                    Required by:
                        unspecified:unspecified:unspecified > org.openrewrite.gradle.tooling:plugin:2.1.0-SNAPSHOT:20240226.180904-8

It does not seem related to my change ...