Copilot commented on code in PR #2502:
URL: https://github.com/apache/groovy/pull/2502#discussion_r3165511139


##########
src/main/java/org/codehaus/groovy/transform/RecordTypeASTTransformation.java:
##########
@@ -183,6 +183,31 @@ public static boolean recordNative(final ClassNode node) {
         return node.getUnresolvedSuperClass() != null && 
RECORD_CLASS_NAME.equals(node.getUnresolvedSuperClass().getName());
     }
 
+    /**
+     * Predicts whether {@code cNode} will be compiled as a native JVM record.
+     * Mirrors the decision logic in {@link #doProcessRecordType} (presence of
+     * {@code @RecordBase}, target bytecode {@code >= 16}, and
+     * {@code @RecordOptions(mode != EMULATE)}) so callers running before the
+     * transform itself fires can act on the same outcome.
+     * <p>
+     * The joint-compilation stub generator uses this at {@code 
Phases.CONVERSION}
+     * to decide whether to emit {@code record Foo(...)} syntax. Unlike
+     * {@link #recordNative(ClassNode)}, which inspects the post-transform
+     * super class, this predicate inspects only the source-level annotations.
+     *
+     * @param cNode          the candidate class node
+     * @param targetBytecode the target bytecode level
+     *                       (see {@link 
CompilerConfiguration#getTargetBytecode()})
+     */
+    @Incubating
+    public static boolean wouldBeNativeRecord(final ClassNode cNode, final 
String targetBytecode) {
+        if (cNode == null || cNode.getAnnotations(MY_TYPE).isEmpty()) return 
false;

Review Comment:
   `wouldBeNativeRecord` calls `StringGroovyMethods.isAtLeast(targetBytecode, 
...)`, which throws if `targetBytecode` is null (it constructs `new 
BigDecimal(left)`). Since the method is public API and callers may not always 
have a target bytecode (and this class already has messaging for the 'unable to 
determine target bytecode' case), add a null/blank guard (return false) before 
calling `isAtLeast`.
   ```suggestion
           if (cNode == null || cNode.getAnnotations(MY_TYPE).isEmpty()) return 
false;
           if (targetBytecode == null || targetBytecode.trim().isEmpty()) 
return false;
   ```



##########
src/test/groovy/org/codehaus/groovy/tools/stubgenerator/RecordTypeJointCompilationStubTest.groovy:
##########
@@ -0,0 +1,64 @@
+/*
+ *  Licensed to the Apache Software Foundation (ASF) under one
+ *  or more contributor license agreements.  See the NOTICE file
+ *  distributed with this work for additional information
+ *  regarding copyright ownership.  The ASF licenses this file
+ *  to you under the Apache License, Version 2.0 (the
+ *  "License"); you may not use this file except in compliance
+ *  with the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ *  Unless required by applicable law or agreed to in writing,
+ *  software distributed under the License is distributed on an
+ *  "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ *  KIND, either express or implied.  See the License for the
+ *  specific language governing permissions and limitations
+ *  under the License.
+ */
+package org.codehaus.groovy.tools.stubgenerator
+
+/**
+ * GROOVY-11974: native records in joint compilation must produce a stub
+ * that javac compiles as a record, so that Java callers see the canonical
+ * constructor and the {@code componentName()} accessors.
+ */
+class RecordTypeJointCompilationStubTest extends StringSourcesStubTestCase {
+
+    @Override
+    protected void configure() {

Review Comment:
   This stub test uses Java record syntax and `Class::isRecord`, which requires 
running on JDK 16+. To keep the test suite green when executed on older JDKs 
(where `javac` can't compile records), add a JUnit assumption (e.g., 
`assumeTrue(isAtLeastJdk('16.0'))`) early in the test (such as in `init()` or 
`configure()`) so it is skipped when records aren't supported.
   ```suggestion
   
       private static boolean isAtLeastJdk(final String version) {
           List<Integer> current = 
System.getProperty('java.specification.version')
               .tokenize('.')
               .collect { it as Integer }
           List<Integer> required = version.tokenize('.')
               .collect { it as Integer }
           int max = Math.max(current.size(), required.size())
           for (int i = 0; i < max; i += 1) {
               int currentPart = (i < current.size() ? current[i] : 0)
               int requiredPart = (i < required.size() ? required[i] : 0)
               if (currentPart != requiredPart) {
                   return currentPart > requiredPart
               }
           }
           true
       }
   
       @Override
       protected void configure() {
           org.junit.Assume.assumeTrue(isAtLeastJdk('16.0'))
   ```



##########
src/test/groovy/org/codehaus/groovy/tools/stubgenerator/RecordTypeJointCompilationStubTest.groovy:
##########
@@ -0,0 +1,64 @@
+/*
+ *  Licensed to the Apache Software Foundation (ASF) under one
+ *  or more contributor license agreements.  See the NOTICE file
+ *  distributed with this work for additional information
+ *  regarding copyright ownership.  The ASF licenses this file
+ *  to you under the Apache License, Version 2.0 (the
+ *  "License"); you may not use this file except in compliance
+ *  with the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ *  Unless required by applicable law or agreed to in writing,
+ *  software distributed under the License is distributed on an
+ *  "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+ *  KIND, either express or implied.  See the License for the
+ *  specific language governing permissions and limitations
+ *  under the License.
+ */
+package org.codehaus.groovy.tools.stubgenerator
+
+/**
+ * GROOVY-11974: native records in joint compilation must produce a stub
+ * that javac compiles as a record, so that Java callers see the canonical
+ * constructor and the {@code componentName()} accessors.
+ */
+class RecordTypeJointCompilationStubTest extends StringSourcesStubTestCase {
+
+    @Override
+    protected void configure() {
+        super.configure()
+        config.targetBytecode = '17'
+    }
+
+    Map<String, String> provideSources() {
+        [
+                'JavaCaller.java': '''
+                    public class JavaCaller {
+                        public static int sumViaRecordAccessors() {
+                            Point p = new Point(3, 4);
+                            return p.x() + p.y();
+                        }
+                        public static String describe() {
+                            return new Point(1, 2).getClass().isRecord()
+                                ? "record" : "class";
+                        }
+                    }
+                ''',
+                'Point.groovy': '''
+                    record Point(int x, int y) {}
+                '''
+        ]
+    }
+
+    void verifyStubs() {
+        def stub = stubJavaSourceFor('Point')
+        // header is rendered as record syntax, not class
+        assert stub =~ /\brecord\s+Point\s*\(\s*int\s+x\s*,\s*int\s+y\s*\)/
+        // no `class Point`, no `extends`
+        assert !(stub =~ /\bclass\s+Point\b/)
+        assert !(stub =~ /\bextends\s+java\.lang\.Object\b/)
+        // no instance-field declarations
+        assert !(stub =~ /(?m)^\s*(?:private|protected|public)?\s*int\s+x\s*;/)

Review Comment:
   The 'no instance-field declarations' assertion only checks for an `int x;` 
field and doesn't cover `y` (or modifiers like `final`). If the stub 
accidentally emitted `private final int x;` or `int y;`, this test would still 
pass. Consider broadening the regex (e.g., allow optional `final` and check 
both component names) to make the test reliably detect any instance field 
emission for record components.
   ```suggestion
           // no instance-field declarations for record components
           assert !(stub =~ 
/(?m)^\s*(?:private|protected|public)?\s*(?:final\s+)?int\s+(?:x|y)\s*;/)
   ```



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to