Guess UTF-16 LE without BOM

app/src/main/java/eu/faircode/email/CharsetHelper.java

@@ -22,6 +22,8 @@ package eu.faircode.email;
 import android.text.TextUtils;
 import android.util.Pair;
 
+import java.io.BufferedInputStream;
+import java.io.IOException;
 import java.nio.ByteBuffer;
 import java.nio.charset.CharacterCodingException;
 import java.nio.charset.Charset;
@@ -130,6 +132,33 @@ public class CharsetHelper {
         return true;
     }
 
+    static Boolean isUTF16LE(BufferedInputStream bis) throws IOException {
+        byte[] bytes = new byte[64];
+        bis.mark(bytes.length);
+        try {
+            int count = bis.read(bytes);
+            if (count < 32)
+                return null;
+
+            int s = ((bytes[0] & 0xff) << 8) | (bytes[1] & 0xff);
+            boolean bom = (s == 0xfeff || s == 0xfffe);
+            if (bom)
+                return null;
+
+            int odd = 0;
+            int even = 0;
+            for (int i = 0; i < count; i++)
+                if (bytes[i] == 0)
+                    if (i % 2 == 0)
+                        even++;
+                    else
+                        odd++;
+            return (even < 30 * count / 100 / 2 && odd > 70 * count / 100 / 2);
+        } finally {
+            bis.reset();
+        }
+    }
+
     static String utf8toW1252(String text) {
         try {
             Charset w1252 = Charset.forName("windows-1252");

app/src/main/java/eu/faircode/email/MessageHelper.java

@@ -3112,7 +3112,21 @@ public class MessageHelper {
                 String result;
 
                 try {
-                    Object content = h.part.getContent();
+                    Object content;
+
+                    // Check for UTF-16 LE without BOM
+                    String pcharset = h.contentType.getParameter("charset");
+                    if ("utf-16".equalsIgnoreCase(pcharset) && override == null) {
+                        String charset = pcharset;
+                        BufferedInputStream bis = new BufferedInputStream(h.part.getDataHandler().getInputStream());
+                        if (Boolean.TRUE.equals(CharsetHelper.isUTF16LE(bis))) {
+                            charset = StandardCharsets.UTF_16LE.name();
+                            Log.e("Charset " + pcharset + " -> " + charset);
+                        }
+                        content = Helper.readStream(bis, Charset.forName(charset));
+                    } else
+                        content = h.part.getContent();
+
                     Log.i("Content class=" + (content == null ? null : content.getClass().getName()));
 
                     if (content == null) {