Fix class-name filter checks for Unicode and long names

utils/src/main/java/datadog/instrument/utils/ClassNameFilter.java

@@ -154,22 +154,22 @@ public static ClassNameFilter readFrom(DataInput in) throws IOException {
    * Computes a 32-bit 'class-code' that includes the length of the package-prefix and simple-name,
    * plus the first and last characters of the simple-name (each truncated to fit into 8-bits.)
    */
-  private static int classCode(String className) {
-    int start = className.lastIndexOf('.') + 1;
-    int end = className.length() - 1;
-    return (0xFF & className.charAt(end)) << 24
-        | (0xFF & className.charAt(start)) << 16
+  private static int classCode(String name) {
+    int start = name.lastIndexOf('.') + 1;
+    int end = name.length() - 1;
+    return (0xFF & name.charAt(end)) << 24
+        | (0xFF & name.charAt(start)) << 16
         | (0xFF & (end - start)) << 8
         | (0xFF & start);
   }
 
   /** Checks whether the 32-bit 'class-code' is consistent with the fully-qualified class-name. */
-  private static boolean checkClassCode(CharSequence className, int code) {
-    int start = (0xFF & code);
-    int end = className.length() - 1;
-    return end - start == (0xFF & (code >> 8))
-        && className.charAt(start) == (0xFF & (code >> 16))
-        && className.charAt(end) == (0xFF & (code >> 24));
+  private static boolean checkClassCode(CharSequence name, int code) {
+    int start = 0xFF & code;
+    int end = name.length() - 1;
+    // Keep the sign bit so a start beyond the end cannot match the encoded length.
+    return (0x800000FF & (end - start)) == (0xFF & (code >> 8))
+        && (((name.charAt(start) ^ code >>> 16) | (name.charAt(end) ^ code >>> 24)) & 0xFF) == 0;
   }
 
   private static int rehash(int oldHash) {

utils/src/test/java/datadog/instrument/utils/ClassNameFilterTest.java

@@ -12,12 +12,52 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.UncheckedIOException;
+import java.util.Arrays;
 import java.util.HashSet;
 import java.util.Set;
 import org.junit.jupiter.api.Test;
+import org.junit.jupiter.params.ParameterizedTest;
+import org.junit.jupiter.params.provider.ValueSource;
 
 class ClassNameFilterTest {
 
+  @ParameterizedTest
+  @ValueSource(strings = {"example.例", "example.例Class", "example.My例"})
+  void unicodeClassNames(String className) {
+    ClassNameFilter filter = new ClassNameFilter(256);
+
+    filter.add(className);
+
+    assertTrue(filter.contains(className));
+  }
+
+  @ParameterizedTest
+  @ValueSource(ints = {256, 257, 512, 513})
+  void longSimpleNames(int length) {
+    char[] simpleName = new char[length];
+    Arrays.fill(simpleName, 'A');
+    String className = "example." + new String(simpleName);
+    ClassNameFilter filter = new ClassNameFilter(256);
+
+    filter.add(className);
+
+    assertTrue(filter.contains(className));
+  }
+
+  @Test
+  void shorterHashCollisionDoesNotReadPastClassName() {
+    char[] padding = new char[248];
+    Arrays.fill(padding, 'A');
+    String className = "a." + new String(padding) + "DVYRQBJ";
+    assertEquals("A".hashCode(), className.hashCode());
+    ClassNameFilter filter = new ClassNameFilter(256);
+
+    filter.add(className);
+
+    assertFalse(filter.contains("A"));
+    assertTrue(filter.contains(className));
+  }
+
   @Test
   void basicOperation() {
     ClassNameFilter filter = new ClassNameFilter(8 * 1024 * 1024);