Use Palantir Java Format

[bric3]

What Does This Do

Switches every Spotless Java configuration from google-java-format 1.36.1 to palantir-java-format 2.97.0 and explicitly selects .style("GOOGLE").
It also updates the contributor and agent guides, including the IntelliJ setup.

Representative Formatting Changes

These examples come from the repository sources reformatted in #12422.

Compact Lambda Pipelines

PluginApplication.java

-        Extension.EXTENSIONS.stream()
-            .filter(ext -> ext.appliesTo(spec))
-            .forEach(
-                ext -> {
+        Extension.EXTENSIONS.stream().filter(ext -> ext.appliesTo(spec)).forEach(ext -> {

Vertically Aligned Fluent Calls

AdviceGeneratorImpl.java

-      final List<Expression> helperTypes =
-          Arrays.stream(helpers)
-              .map(type -> new StringLiteralExpr(type.getClassName()))
-              .collect(Collectors.toList());
+      final List<Expression> helperTypes = Arrays.stream(helpers)
+          .map(type -> new StringLiteralExpr(type.getClassName()))
+          .collect(Collectors.toList());

Nested Callbacks Without an Indentation Pyramid

CoreTracer.java

-    sharedCommunicationObjects.whenReady(
-        () ->
-            AgentTaskScheduler.get()
-                .execute(
-                    () -> {
-                      startMetricsAggregation(config, sco);
-                      maybeStartLogsExport(config);
-                    }));
+    sharedCommunicationObjects.whenReady(() -> AgentTaskScheduler.get().execute(() -> {
+      startMetricsAggregation(config, sco);
+      maybeStartLogsExport(config);
+    }));

Nested Lambda Pipelines Without Stair-Step Indentation

AssertBuilder.java

     return type.getAnnotationByName("AutoService")
-        .<Set<Class<?>>>map(
-            annotation ->
-                annotation.asNormalAnnotationExpr().getPairs().stream()
-                    .filter(pair -> pair.getNameAsString().equals("value"))
-                    .flatMap(
-                        pair ->
-                            pair.getValue().asArrayInitializerExpr().getValues().stream()
-                                .map(
-                                    value ->
-                                        value
-                                            .asClassExpr()
-                                            .getType()
-                                            .resolve()
-                                            .asReferenceType()
-                                            .getTypeDeclaration()
-                                            .get()
-                                            .getQualifiedName()))
-                    .map(AssertBuilder::loadClass)
-                    .collect(Collectors.toSet()))
+        .<Set<Class<?>>>map(annotation -> annotation.asNormalAnnotationExpr().getPairs().stream()
+            .filter(pair -> pair.getNameAsString().equals("value"))
+            .flatMap(pair -> pair.getValue().asArrayInitializerExpr().getValues().stream()
+                .map(value -> value
+                    .asClassExpr()
+                    .getType()
+                    .resolve()
+                    .asReferenceType()
+                    .getTypeDeclaration()
+                    .get()
+                    .getQualifiedName()))
+            .map(AssertBuilder::loadClass)
+            .collect(Collectors.toSet()))
         .orElse(Collections.emptySet());

Switch Blocks Attached to Their Labels

JsonToExpressionConverter.java

-      case "not":
-        {
-          JsonReader.Token token = reader.peek();
-          if (token == BEGIN_ARRAY || token == STRING || token == NUMBER) {
-            throw new UnsupportedOperationException(
-                "Operation 'not' expects a predicate as its argument");
-          }
-          return DSL.not(createPredicate(reader));
-        }
+      case "not": {
+        JsonReader.Token token = reader.peek();
+        if (token == BEGIN_ARRAY || token == STRING || token == NUMBER) {
+          throw new UnsupportedOperationException(
+              "Operation 'not' expects a predicate as its argument");
+        }
+        return DSL.not(createPredicate(reader));
+      }

Motivation

The experiment deliberately selects GOOGLE style to preserve the repository's existing two-space indentation and 100-character line width. This keeps the evaluation focused on Palantir's lambda and fluent-chain layout instead of combining it with a repository-wide indentation and line-width change. It also reduces the stacked reformat from 6,133 files with the default PALANTIR style to 1,855 files.

Additional Notes

This is the base of a two-PR experiment. spotlessCheck is expected to report Java format violations on this PR alone; the stacked reformatting PR #12422 makes the new formatter idempotent.

The Palantir IntelliJ plugin currently exposes only the default PALANTIR style, so the contributor guide keeps the Spotless Gradle tasks as the formatting source of truth for this GOOGLE-style experiment.

Palantir 2.97.0 with GOOGLE style keeps the closing parenthesis beside the final parameter:

private static Function<CallSiteSpecification, CallSiteResult> generateAdviceClosure(
    final Configuration configuration) {

Spotless exposes Palantir style and Javadoc options, but no independent indentation or closing-parenthesis option.

Contributor Checklist

• Format the title according to the contribution guidelines
• Assign the type: and (comp: or inst:) labels in addition to any other useful labels
• Avoid using close, fix, or any linking keywords when referencing an issue
  Use solves instead, and assign the PR milestone to the issue
• Update the CODEOWNERS file on source file addition, migration, or deletion
• Update public documentation with any new configuration flags or behaviors
• Once approved, use merge queue to merge the PR