Improved metadata trimming, to better handle multibyte UTF-8 characters.

src/main/java/org/qortal/arbitrary/metadata/ArbitraryDataTransactionMetadata.java

@@ -7,6 +7,7 @@ import org.qortal.arbitrary.misc.Category;
 import org.qortal.repository.DataException;
 import org.qortal.utils.Base58;
 
+import java.nio.charset.StandardCharsets;
 import java.nio.file.Path;
 import java.util.ArrayList;
 import java.util.Arrays;
@@ -217,6 +218,25 @@ public class ArbitraryDataTransactionMetadata extends ArbitraryDataMetadata {
 
     // Static helper methods
 
+    public static String trimUTF8String(String string, int maxLength) {
+        byte[] inputBytes = string.getBytes(StandardCharsets.UTF_8);
+        int length = Math.min(inputBytes.length, maxLength);
+        byte[] outputBytes = new byte[length];
+
+        System.arraycopy(inputBytes, 0, outputBytes, 0, length);
+        String result = new String(outputBytes, StandardCharsets.UTF_8);
+
+        // check if last character is truncated
+        int lastIndex = result.length() - 1;
+
+        if (lastIndex > 0 && result.charAt(lastIndex) != string.charAt(lastIndex)) {
+            // last character is truncated so remove the last character
+            return result.substring(0, lastIndex);
+        }
+
+        return result;
+    }
+
     public static String limitTitle(String title) {
         if (title == null) {
             return null;
@@ -225,7 +245,7 @@ public class ArbitraryDataTransactionMetadata extends ArbitraryDataMetadata {
             return null;
         }
 
-        return title.substring(0, Math.min(title.length(), MAX_TITLE_LENGTH));
+        return trimUTF8String(title, MAX_TITLE_LENGTH);
     }
 
     public static String limitDescription(String description) {
@@ -236,7 +256,7 @@ public class ArbitraryDataTransactionMetadata extends ArbitraryDataMetadata {
             return null;
         }
 
-        return description.substring(0, Math.min(description.length(), MAX_DESCRIPTION_LENGTH));
+        return trimUTF8String(description, MAX_DESCRIPTION_LENGTH);
     }
 
     public static List<String> limitTags(List<String> tags) {

src/test/java/org/qortal/test/arbitrary/ArbitraryTransactionMetadataTests.java

@@ -248,6 +248,47 @@ public class ArbitraryTransactionMetadataTests extends Common {
         }
     }
 
+    @Test
+    public void testUTF8Metadata() throws DataException, IOException, MissingDataException {
+        try (final Repository repository = RepositoryManager.getRepository()) {
+            PrivateKeyAccount alice = Common.getTestAccount(repository, "alice");
+            String publicKey58 = Base58.encode(alice.getPublicKey());
+            String name = "TEST"; // Can be anything for this test
+            String identifier = null; // Not used for this test
+            Service service = Service.ARBITRARY_DATA;
+            int chunkSize = 100;
+            int dataLength = 900; // Actual data length will be longer due to encryption
+
+            // Example (modified) strings from real world content
+            String title = "Доля юаня в трансграничных Доля юаня в трансграничных";
+            String description = "Когда рыночек порешал";
+            List<String> tags = Arrays.asList("Доля", "юаня", "трансграничных");
+            Category category = Category.OTHER;
+
+            // Register the name to Alice
+            RegisterNameTransactionData transactionData = new RegisterNameTransactionData(TestTransaction.generateBase(alice), name, "");
+            transactionData.setFee(new RegisterNameTransaction(null, null).getUnitFee(transactionData.getTimestamp()));
+            TransactionUtils.signAndMint(repository, transactionData, alice);
+
+            // Create PUT transaction
+            Path path1 = ArbitraryUtils.generateRandomDataPath(dataLength);
+            ArbitraryDataFile arbitraryDataFile = ArbitraryUtils.createAndMintTxn(repository, publicKey58, path1, name,
+                    identifier, ArbitraryTransactionData.Method.PUT, service, alice, chunkSize, 0L, true,
+                    title, description, tags, category);
+
+            // Check the chunk count is correct
+            assertEquals(10, arbitraryDataFile.chunkCount());
+
+            // Check the metadata is correct
+            String expectedTrimmedTitle = "Доля юаня в трансграничных Доля юаня в тран";
+            assertEquals(expectedTrimmedTitle, arbitraryDataFile.getMetadata().getTitle());
+            assertEquals(description, arbitraryDataFile.getMetadata().getDescription());
+            assertEquals(tags, arbitraryDataFile.getMetadata().getTags());
+            assertEquals(category, arbitraryDataFile.getMetadata().getCategory());
+            assertEquals("text/plain", arbitraryDataFile.getMetadata().getMimeType());
+        }
+    }
+
     @Test
     public void testMetadataLengths() throws DataException, IOException, MissingDataException {
         try (final Repository repository = RepositoryManager.getRepository()) {