Fix SymDB when processing corrupted jars (#10453)

when processing jar file that are corrupted exceptions can be raised
and will stop the process. Make sure we can process the next jars.

Author: jpbempel

dd-java-agent/agent-debugger/src/main/java/com/datadog/debugger/symbol/SymDBEnablement.java

@@ -143,7 +143,8 @@ private void extractSymbolForLoadedClasses(SymDBReport symDBReport) {
       try {
         jarPath = JarScanner.extractJarPath(clazz, symDBReport);
       } catch (URISyntaxException e) {
-        throw new RuntimeException(e);
+        LOGGER.debug("Failed to extract jar path for class {}", clazz.getTypeName(), e);
+        continue;
       }
       if (jarPath == null) {
         continue;
@@ -152,7 +153,11 @@ private void extractSymbolForLoadedClasses(SymDBReport symDBReport) {
         symDBReport.addMissingJar(jarPath.toString());
         continue;
       }
-      symbolAggregator.scanJar(symDBReport, jarPath, baos, buffer);
+      try {
+        symbolAggregator.scanJar(symDBReport, jarPath, baos, buffer);
+      } catch (Exception ex) {
+        LOGGER.debug("Failed to scan jar {}", jarPath, ex);
+      }
     }
   }
 }

dd-java-agent/agent-debugger/src/main/java/com/datadog/debugger/symbol/SymbolAggregator.java

@@ -158,7 +158,11 @@ void scanQueuedJars(SymbolAggregator symbolAggregator) {
     while (!jarsToScanQueue.isEmpty()) {
       String jarPath = jarsToScanQueue.poll();
       LOGGER.debug("Scanning queued jar: {}", jarPath);
-      scanJar(SymDBReport.NO_OP, Paths.get(jarPath), baos, buffer);
+      try {
+        scanJar(SymDBReport.NO_OP, Paths.get(jarPath), baos, buffer);
+      } catch (Exception ex) {
+        LOGGER.debug("Failed to scan jar {}", jarPath, ex);
+      }
     }
   }
 

dd-java-agent/agent-debugger/src/test/java/com/datadog/debugger/symbol/SymDBEnablementTest.java

@@ -25,9 +25,11 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.lang.instrument.Instrumentation;
+import java.net.MalformedURLException;
 import java.net.URISyntaxException;
 import java.net.URL;
 import java.net.URLClassLoader;
+import java.nio.file.Path;
 import java.util.Collections;
 import java.util.jar.JarEntry;
 import java.util.jar.JarFile;
@@ -111,11 +113,7 @@ public void noIncludesFilterOutDatadogClass() {
 
   @Test
   public void parseLoadedClass() throws ClassNotFoundException, IOException {
-    final String CLASS_NAME = "com.datadog.debugger.symbol.SymbolExtraction01";
-    URL jarFileUrl = getClass().getResource("/debugger-symbol.jar");
-    URL jarUrl = new URL("jar:file:" + jarFileUrl.getFile() + "!/");
-    URLClassLoader urlClassLoader = new URLClassLoader(new URL[] {jarUrl}, null);
-    Class<?> testClass = urlClassLoader.loadClass(CLASS_NAME);
+    Class<?> testClass = loadSymbolClassFromJar();
     when(instr.getAllLoadedClasses()).thenReturn(new Class[] {testClass});
     when(config.getThirdPartyIncludes())
         .thenReturn(
@@ -130,14 +128,50 @@ public void parseLoadedClass() throws ClassNotFoundException, IOException {
     verify(instr).addTransformer(any(SymbolExtractionTransformer.class));
     ArgumentCaptor<String> captor = ArgumentCaptor.forClass(String.class);
     verify(symbolAggregator, times(2))
-        .parseClass(any(), captor.capture(), any(), eq(jarFileUrl.getFile()));
+        .parseClass(
+            any(),
+            captor.capture(),
+            any(),
+            eq(getClass().getResource("/debugger-symbol.jar").getFile()));
     assertEquals(
         "com/datadog/debugger/symbol/SymbolExtraction01.class", captor.getAllValues().get(0));
     assertEquals(
         "BOOT-INF/classes/org/springframework/samples/petclinic/vet/VetController.class",
         captor.getAllValues().get(1));
   }
 
+  @Test
+  public void processCorruptedJar() throws ClassNotFoundException, MalformedURLException {
+    Class<?> testClass = loadSymbolClassFromJar();
+    when(instr.getAllLoadedClasses())
+        .thenReturn(new Class[] {SymDBEnablementTest.class, testClass});
+    ClassNameFiltering classNameFiltering = ClassNameFiltering.allowAll();
+    SymbolAggregator symbolAggregatorMock = mock(SymbolAggregator.class);
+    doAnswer(
+            invocation -> {
+              Path arg = invocation.getArgument(1, Path.class);
+              if (arg.toString().endsWith("/debugger-symbol.jar")) {
+                return null;
+              }
+              throw new IOException("Corrupted jar");
+            })
+        .when(symbolAggregatorMock)
+        .scanJar(any(), any(), any(), any());
+    SymDBEnablement symDBEnablement =
+        new SymDBEnablement(instr, config, symbolAggregatorMock, classNameFiltering);
+    symDBEnablement.startSymbolExtraction();
+    verify(symbolAggregatorMock, times(2)).scanJar(any(), any(), any(), any());
+  }
+
+  private Class<?> loadSymbolClassFromJar() throws MalformedURLException, ClassNotFoundException {
+    final String CLASS_NAME = "com.datadog.debugger.symbol.SymbolExtraction01";
+    URL jarFileUrl = getClass().getResource("/debugger-symbol.jar");
+    URL jarUrl = new URL("jar:file:" + jarFileUrl.getFile() + "!/");
+    URLClassLoader urlClassLoader = new URLClassLoader(new URL[] {jarUrl}, null);
+    Class<?> testClass = urlClassLoader.loadClass(CLASS_NAME);
+    return testClass;
+  }
+
   @Test
   public void parseLoadedClassFromDirectory()
       throws ClassNotFoundException, IOException, URISyntaxException {

dd-java-agent/agent-debugger/src/test/java/com/datadog/debugger/symbol/SymbolAggregatorTest.java

@@ -46,6 +46,36 @@ void testScanQueuedJars() {
         captor.getAllValues().get(2));
   }
 
+  @Test
+  void testScanQueuedCorruptedJars() {
+    SymbolSink symbolSink = mock(SymbolSink.class);
+    SymbolAggregator symbolAggregator =
+        spy(new SymbolAggregator(ClassNameFiltering.allowAll(), emptyList(), symbolSink, 1));
+    // add first a corrupted jar
+    URL corruptedUrl = getClass().getResource("/com/datadog/debugger/classfiles/CommandLine.class");
+    CodeSource corruptedCodeSource =
+        new CodeSource(corruptedUrl, (java.security.cert.Certificate[]) null);
+    ProtectionDomain corruptedProtectionDomain = new ProtectionDomain(corruptedCodeSource, null);
+    symbolAggregator.parseClass(null, null, corruptedProtectionDomain);
+    // add second a clean jar
+    URL jarFileUrl = getClass().getResource("/debugger-symbol.jar");
+    CodeSource codeSource = new CodeSource(jarFileUrl, (java.security.cert.Certificate[]) null);
+    ProtectionDomain protectionDomain = new ProtectionDomain(codeSource, null);
+    symbolAggregator.parseClass(null, null, protectionDomain);
+    symbolAggregator.scanQueuedJars(null);
+    // clean jar should have been processed
+    ArgumentCaptor<String> captor = ArgumentCaptor.forClass(String.class);
+    verify(symbolAggregator, atLeastOnce())
+        .parseClass(any(), captor.capture(), any(), eq(jarFileUrl.getFile()));
+    // captor.getAllValues().get(0) is the first argument of the first invocation of parseClass with
+    // null
+    assertEquals(
+        "com/datadog/debugger/symbol/SymbolExtraction01.class", captor.getAllValues().get(1));
+    assertEquals(
+        "BOOT-INF/classes/org/springframework/samples/petclinic/vet/VetController.class",
+        captor.getAllValues().get(2));
+  }
+
   @Test
   @DisabledIf(
       value = "datadog.environment.JavaVirtualMachine#isJ9",