gnodet-bot commented on code in PR #12694: URL: https://github.com/apache/maven/pull/12694#discussion_r4057964158
########## impl/maven-logging/src/test/java/org/apache/maven/slf4j/MavenJulHandlerTest.java: ########## @@ -0,0 +1,150 @@ +/* + * 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.apache.maven.slf4j; + +import java.lang.reflect.InvocationTargetException; +import java.lang.reflect.Method; +import java.util.logging.Level; + +import org.junit.jupiter.api.Test; +import org.junit.jupiter.params.ParameterizedTest; +import org.junit.jupiter.params.provider.CsvSource; +import org.slf4j.spi.LocationAwareLogger; + +import static org.junit.jupiter.api.Assertions.assertEquals; +import static org.junit.jupiter.api.Assertions.assertNull; + +/** + * Tests for {@link MavenJulHandler}, focused on the JUL→SLF4J level + * mapping and metadata lifecycle. + */ +class MavenJulHandlerTest { + + /** + * Table test for JUL level → SLF4J level mapping. + * Verifies the mapping documented in the class Javadoc. + */ + @ParameterizedTest(name = "JUL {0} -> SLF4J level {1}") + @CsvSource({ + "FINEST, 0", // TRACE_INT = 0 + "FINER, 10", // DEBUG_INT = 10 + "FINE, 10", // DEBUG_INT = 10 + "CONFIG, 20", // INFO_INT = 20 + "INFO, 20", // INFO_INT = 20 + "WARNING, 30", // WARN_INT = 30 + "SEVERE, 40", // ERROR_INT = 40 + }) + void julLevelMapsToCorrectSlf4jLevel(String julLevelName, int expectedSlf4jLevel) throws Exception { + Level julLevel = Level.parse(julLevelName); + int actual = invokeJulLevelToSlf4j(julLevel); + assertEquals( + expectedSlf4jLevel, actual, "JUL " + julLevelName + " should map to SLF4J level " + expectedSlf4jLevel); + } + + /** + * Verify that FINEST maps to TRACE (not DEBUG) — this is the key + * distinction for the TRACE/DEBUG separation. + */ + @Test + void finestMapsToTrace() throws Exception { + assertEquals( + LocationAwareLogger.TRACE_INT, + invokeJulLevelToSlf4j(Level.FINEST), + "FINEST should map to TRACE, not DEBUG"); + } + + /** + * Verify that CONFIG maps to INFO (not DEBUG) — CONFIG is JUL's + * informational level for static configuration, not a debug level. + */ + @Test + void configMapsToInfo() throws Exception { + assertEquals(LocationAwareLogger.INFO_INT, invokeJulLevelToSlf4j(Level.CONFIG), "CONFIG should map to INFO"); + } + + /** + * Verify that JUL metadata is null when no log event is being processed. + */ + @Test + void julMetadataIsNullOutsidePublish() { + assertNull(MavenJulHandler.getJulMetadata(), "JUL metadata should be null outside a publish() call"); + } + + /** + * Verify that a recursive call to {@link MavenJulHandler#publish} from + * within a {@code publish()} call on the same thread is silently dropped + * instead of crashing with {@code IllegalStateException("Recursive update")}. + * <p> + * This is the reentrancy scenario that occurs when JLine's + * {@code StyleResolver} logs a JUL DEBUG event while + * {@code MavenSimpleLogger.renderLevel()} is lazily initialising ANSI + * colour strings during terminal construction. + */ + @Test + void publishIsReentrantSafe() throws Exception { + MavenJulHandler handler = new MavenJulHandler(); + java.util.logging.LogRecord outerRecord = new java.util.logging.LogRecord(Level.INFO, "outer"); + + // Simulate re-entrance: set IN_PUBLISH=true before calling publish(). + // The reentrancy guard should short-circuit without throwing or logging. + java.util.logging.Logger julLogger = + java.util.logging.LogManager.getLogManager().getLogger(""); + java.util.logging.Handler[] saved = julLogger.getHandlers(); + for (java.util.logging.Handler h : saved) { + julLogger.removeHandler(h); + } + + // The test verifies that publish() does not throw. + // We can't easily simulate the full SLF4J pipeline here, so we just + // call publish() with a null-logger-name record (which returns early + // before reaching SLF4J) after setting IN_PUBLISH to true, verifying + // the guard works. + java.lang.reflect.Field inPublishField = MavenJulHandler.class.getDeclaredField("IN_PUBLISH"); + inPublishField.setAccessible(true); + @SuppressWarnings("unchecked") + ThreadLocal<Boolean> inPublish = (ThreadLocal<Boolean>) inPublishField.get(null); + + // Simulate being inside publish() + inPublish.set(Boolean.TRUE); + try { + // A nested call should be dropped without throwing + handler.publish(outerRecord); + // If we reach here, the guard worked correctly + } finally { + inPublish.remove(); + // Restore handlers + for (java.util.logging.Handler h : saved) { + julLogger.addHandler(h); + } + } Review Comment: **[low] `publishIsReentrantSafe()` simulates reentrancy via reflection but doesn't exercise the real crash path** The test accesses `IN_PUBLISH` via `getDeclaredField("IN_PUBLISH")` and pre-sets it to `true`, then calls `publish()` which returns early at the reentrancy check. This verifies the guard's early-return branch, but not the *actual crash scenario*: `MavenSimpleLogger.renderLevel()` calling `StyleResolver.resolve()` → `ConcurrentHashMap.computeIfAbsent` reentrancy. Using reflection ties the test to the private field name and ThreadLocal type — renaming `IN_PUBLISH` silently breaks the test with an `NoSuchFieldException` instead of a compile error. The test also removes JUL root handlers unnecessarily (they aren't involved in the guard check path). A simpler and more honest test would install `MavenJulHandler` as the only handler, then fire a JUL event from within an SLF4J Logger callback that itself fires a JUL event — but that requires a real SLF4J binding. Alternatively, expose a package-private `setReentrancyGuardForTest()` test hook so the test doesn't need reflection. This is low priority given the production fix is correct, but the test's coverage claim in its Javadoc is stronger than what it actually exercises. -- 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]
