refactor: migrate Xtend to Java - com.avaloq.tools.ddk.xtext.export

[joaodinissf]

Summary

Migrates the last legacy Xtend source of com.avaloq.tools.ddk.xtext.export to Java 21: generator/ExportGeneratorX — the naming, prefix, supertype/interface lookup and user-data helpers the export inferrer uses.

The module keeps its Xbase layer in jvmmodel/*.xtend, so the Xtend build infrastructure (xtend-gen, .classpath, build.properties) is untouched.

Stacked on the xtext.expression migration (this class injects its Naming).

Public API is unchanged. The sole consumer is ExportJvmModelInferrer.xtend, which binds through @Inject extension — allUserData(...), constantName(...), typeMap(...), sortedExportsByEPackage(...), getSuperInterfaces(...), exportForType(...) and the get* provider-name helpers. It compiles unchanged against the Java. (superType and getName(ExportModel) are reached only from inside this class; model.name in the inferrer resolves to ExportModel's own EMF getter, which takes priority over a same-named extension.)

Faithfulness notes

• Dispatcher allUserData(Export) keeps the compiler's shape: protected _allUserData(Export) / _allUserData(Void), non-null branch first, Void sentinel second, no terminal exception (as in the compiler output); the recursion through the public dispatcher is preserved.
• constantName(EAttribute, EClass) / constantName(UserData, EClass) and getResourceDescriptionManager(ExportModel) / (Grammar) stay plain overloads.
• getGrammar keeps IterableExtensions.head semantics via Guava Iterables.getFirst(contents, null) (Guava is already a dependency through ListMultimap): empty contents yield null, the vestigial ?. guard is preserved.
• exportForType uses Stream.filter(...).findFirst(), matching IterableExtensions.findFirst. getInterfacesForType collects eagerly, matching the compiler output: Interface.getType() resolves EMF proxies in place, so evaluating the predicate over every declaration is an observable side effect rather than wasted work.
• Guava Function lambdas → Export::getType method references; newArrayList() → new ArrayList<>() (results stay mutable, as the callers rely on). Xtend == → Objects.equals, matching the compiler output.
• Javadoc preserved verbatim; only @param/@return tags and two sentence-ending periods were added to Javadoc that already existed, because Checkstyle enforces them on .java and never saw the .xtend. No Javadoc was invented for undocumented methods. The literal 3 in getPrefix became the named constant FIRST_PACKAGE_SEGMENT.

Known deltas

• The Xtend compiler stamped @XbaseGenerated on the generated dispatchers; the migrated Java does not. The annotation is runtime-retained on public methods in an exported package, so it is a reflectively observable delta — but nothing in the target platform reads it (only JvmModelGenerator emits it), and re-applying a "machine-generated" marker to hand-written code would be false.

Verification

• Ground truth: fresh xtend-gen from a clean build of the base commit.
• Two independent translations were produced and reconciled against the compiler output. An independent review of the reconciled result then caught one fidelity-relevant difference — getInterfacesForType had been left short-circuiting where the compiler output evaluates the predicate over every declaration — which is fixed above; the other translation had it right. Remaining differences between the two were style.
• Bytecode comparison against the compiler's classes: public/protected API identical; ordered string constants per method identical; dispatcher branch order identical. Only private members differ (the removed compiler lambdas).
• Gates: -amd compile (rebuilds the Xtend caller against the Java) ✅; pmd:check checkstyle:check spotbugs:check ✅; full reactor clean verify with the CI static-analysis goals ✅ (354 tests, 0 failures, 2 skipped). No version bump needed: the bundle is already at 17.3.3, ahead of the baseline.

🤖 Generated with Claude Code