Merge pull request #119 from bgrozev/edge-workaround

fix: Works around a bug in Edge

Author: bbaldino

src/main/java/org/ice4j/attribute/UsernameAttribute.java

@@ -17,6 +17,8 @@
  */
 package org.ice4j.attribute;
 
+import org.ice4j.*;
+
 import java.util.*;
 
 /**
@@ -33,6 +35,24 @@ public class UsernameAttribute extends Attribute
      */
     public static final String NAME = "USERNAME";
 
+    /**
+     * The name of the property which controls whether trailing zeroes should
+     * be stripped when decoding USERNAME attributes.
+     */
+    public static final String STRIP_TRAILING_ZEROES_PNAME
+        = "org.ice4j.attribute.UsernameAttribute.STRIP_TRAILING_ZEROES";
+
+    /**
+     * Whether trailing zeroes should be stripped when decoding USERNAME
+     * attributes. This is necessary to work around a bug in Edge described
+     * here:
+     * https://developer.microsoft.com/en-us/microsoft-edge/platform/issues/12332457/
+     * and here:
+     * https://github.com/jitsi/lib-jitsi-meet/issues/498
+     */
+    private boolean stripTrailingZeroes
+        = StackProperties.getBoolean(STRIP_TRAILING_ZEROES_PNAME, false);
+
     /**
      * Username value.
      */
@@ -56,8 +76,17 @@ public class UsernameAttribute extends Attribute
      *   offset is equal to the index of the first byte after length)
      * @param length the length of the binary array.
      */
+    @Override
     void decodeAttributeBody(byte[] attributeValue, char offset, char length)
     {
+        if (stripTrailingZeroes)
+        {
+            while (length > 0 && attributeValue[offset + length] == 0)
+            {
+                length--;
+            }
+        }
+
         username = new byte[length];
         System.arraycopy(attributeValue, offset, username, 0, length);
     }